0% found this document useful (0 votes)

632 views471 pages

Reference Book

Uploaded by

Marwa Fourati Ep Hentati

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

632 views471 pages

Reference Book

Uploaded by

Marwa Fourati Ep Hentati

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 471

Trimble

eCognition Developer
Reference Book

Trimble Documentation:

eCognition Developer 9.0

Reference Book
Imprint and Version
Document Version 9.0.1
Copyright 2014 Trimble Germany GmbH. All rights reserved. This document may be copied and printed only in accordance with the terms of the Frame
License Agreement for End Users of the related eCognition software.
Published by:
Trimble Germany GmbH, Arnulfstrasse 126, D-80636 Munich, Germany
.
Phone: +49898905710 .. Fax: +4989890571411
Web: www.eCognition.com
Dear User,
Thank you for using eCognition software. We appreciate being of service to
you with image analysis solutions. At Trimble we constantly strive to improve
our products. We therefore appreciate all comments and suggestions for improvements concerning our software, training, and documentation. Feel free to
contact us via the web form on www.eCognition.com/support. Thank you.
Legal Notes
Trimble and eCognition are registered trademarks of Trimble Germany
GmbH in Germany and other countries. All other product names, company
names, and brand names mentioned in this document may be trademark properties of their respective holders.
Protected by patents EP0858051; WO0145033; WO2004036337; US
6,832,002; US 7,437,004; US 7,574,053 B2; US 7,146,380; US 7,467,159 B;
US 7,873,223; US 7,801,361 B2.
Acknowledgments
Portions of this product are based in part on third-party software components:
eCognition Developer 2014 Trimble Germany GmbH, Arnulfstrasse 126,
80636 Munich, Germany. All rights reserved.
The Visualisation Toolkit (VTK) 19932006 Ken Martin, Will Schroeder,
Bill Lorensen. All rights reserved.
Insight Segmentation and Registration Toolkit (ITK) 1999- 2003 Insight Software Consortium. All rights reserved.
*

Typeset by Wikipublisher

All rights reserved.

2014 Trimble Documentation, Mnchen, Germany

Day of print: 11 June 2014

Contents
1 Introduction
1.1 Symbols and Expressions . . . . . . . . . . . . . . .
1.1.1 Basic Mathematical Notations . . . . . . . .
1.1.2 Image Layer and Scene . . . . . . . . . . . .
1.1.3 Region . . . . . . . . . . . . . . . . . . . .
1.1.4 Pixel Set . . . . . . . . . . . . . . . . . . .
1.1.5 Image Object . . . . . . . . . . . . . . . . .
1.1.6 Image Objects Hierarchy . . . . . . . . . . .
1.1.7 Class-Related Set . . . . . . . . . . . . . . .
1.2 Coordinate Systems Used in eCognition Software . .
1.2.1 Pixel Coordinate System . . . . . . . . . . .
1.2.2 User Coordinate System . . . . . . . . . . .
1.3 Image Layer Related Features . . . . . . . . . . . .
1.3.1 Scene . . . . . . . . . . . . . . . . . . . . .
1.3.2 Maps . . . . . . . . . . . . . . . . . . . . .
1.3.3 Region . . . . . . . . . . . . . . . . . . . .
1.3.4 Image Layer . . . . . . . . . . . . . . . . .
1.3.5 Image Layer Intensity on Pixel Sets . . . . .
1.4 Image Object Related Features . . . . . . . . . . . .
1.4.1 Image Object . . . . . . . . . . . . . . . . .
1.4.2 Image Object Hierarchy . . . . . . . . . . .
1.5 Class-Related Features . . . . . . . . . . . . . . . .
1.5.1 Class-Related Sets . . . . . . . . . . . . . .
1.6 Shape-Related Features . . . . . . . . . . . . . . . .
1.6.1 Parameters . . . . . . . . . . . . . . . . . .
1.6.2 Expression . . . . . . . . . . . . . . . . . .
1.6.3 Shape Approximations Based on Eigenvalues
1.6.4 Elliptic Approximation . . . . . . . . . . . .
1.7 About Algorithms . . . . . . . . . . . . . . . . . . .
1.7.1 Creating a Process . . . . . . . . . . . . . .
1.7.2 Specifying Algorithm Parameters . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

1
2
2
2
3
3
3
4
4
5
5
6
7
7
8
9
9
10
11
11
15
17
17
18
18
18
18
19
20
20
20

2 Process-Related Operation Algorithms

2.1 Execute Child Processes . . . . . .
2.1.1 Supported Domains . . . . .
2.2 Execute Child As Series . . . . . .
2.2.1 Supported Domains . . . . .
2.2.2 Algorithm Parameters . . .
2.3 If, Then and Else . . . . . . . . . .

.
.
.
.
.
.

22
22
22
24
24
24
24

.
.
.
.
.
.

CONTENTS
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

25
25
25
25
25
25
25
26

3 Segmentation Algorithms
3.1 Chessboard Segmentation . . . . . . . . . . . .
3.1.1 Supported Domains . . . . . . . . . . .
3.1.2 Algorithm Parameters . . . . . . . . .
3.2 Quadtree-Based Segmentation . . . . . . . . .
3.2.1 Supported Domains . . . . . . . . . . .
3.2.2 Algorithm Parameters . . . . . . . . .
3.2.3 Thematic Layer Weights . . . . . . . .
3.3 Contrast Split Segmentation . . . . . . . . . .
3.3.1 Supported Domains . . . . . . . . . . .
3.3.2 Settings . . . . . . . . . . . . . . . . .
3.3.3 Advanced Settings . . . . . . . . . . .
3.4 Multiresolution Segmentation . . . . . . . . .
3.4.1 Supported Domains . . . . . . . . . . .
3.4.2 Level Settings . . . . . . . . . . . . . .
3.4.3 Segmentation Settings . . . . . . . . .
3.4.4 Composition of Homogeneity Criterion
3.5 Spectral Difference Segmentation . . . . . . .
3.5.1 Supported Domains . . . . . . . . . . .
3.5.2 Level Settings . . . . . . . . . . . . . .
3.5.3 Segmentation Settings . . . . . . . . .
3.6 Multi-Threshold Segmentation . . . . . . . . .
3.6.1 Supported Domains . . . . . . . . . . .
3.6.2 Level Settings . . . . . . . . . . . . . .
3.7 Contrast Filter Segmentation . . . . . . . . . .
3.7.1 Supported Domains . . . . . . . . . . .
3.7.2 Chessboard Settings . . . . . . . . . .
3.7.3 Input Parameters . . . . . . . . . . . .
3.7.4 Shape Criteria Settings . . . . . . . . .
3.7.5 Classication Parameters . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

28
28
29
29
29
30
30
32
32
32
32
34
35
37
37
38
39
40
40
40
41
41
42
42
43
43
43
44
45
45

4 Basic Classification Algorithms

4.1 Assign Class . . . . . . . . . . . . .
4.1.1 Use Class . . . . . . . . . .
4.2 Classication . . . . . . . . . . . .
4.2.1 Active Classes . . . . . . .
4.3 Hierarchical Classication . . . . .
4.3.1 Active Classes . . . . . . .
4.3.2 Use Class-Related Features .
4.4 Remove Classication . . . . . . .
4.4.1 Classes . . . . . . . . . . .

.
.
.
.
.
.
.
.
.

47
47
47
47
47
48
48
48
48
48

2.4
2.5
2.6

2.3.1 Supported Domains . .

Throw . . . . . . . . . . . . .
2.4.1 Supported Domains . .
Catch . . . . . . . . . . . . .
2.5.1 Supported Domains . .
Set Rule Set Options . . . . .
2.6.1 Supported Domains . .
2.6.2 Algorithm Parameters

iii
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

Reference Book

11 June 2014

CONTENTS

5 Advanced Classification Algorithms

5.1 Find Domain Extrema . . . . . . . . . . . . . . . . . . . .
5.1.1 Supported Domains . . . . . . . . . . . . . . . . .
5.1.2 Extrema Settings . . . . . . . . . . . . . . . . . .
5.1.3 Classication Settings . . . . . . . . . . . . . . .
5.2 Find Local Extrema . . . . . . . . . . . . . . . . . . . . .
5.2.1 Supported Domains . . . . . . . . . . . . . . . . .
5.2.2 Search Settings . . . . . . . . . . . . . . . . . . .
5.2.3 Conditions . . . . . . . . . . . . . . . . . . . . .
5.2.4 Classication Settings . . . . . . . . . . . . . . .
5.3 Find Enclosed by Class . . . . . . . . . . . . . . . . . . .
5.3.1 Supported Domains . . . . . . . . . . . . . . . . .
5.3.2 Search Settings . . . . . . . . . . . . . . . . . . .
5.3.3 Classication Settings . . . . . . . . . . . . . . .
5.4 Find Enclosed by Image Object . . . . . . . . . . . . . . .
5.4.1 Classication Settings . . . . . . . . . . . . . . .
5.5 Connector . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5.1 Connector Settings . . . . . . . . . . . . . . . . .
5.5.2 Classication Settings . . . . . . . . . . . . . . .
5.6 Assign Class By Slice Overlap (Prototype) . . . . . . . . .
5.6.1 Supported Domains . . . . . . . . . . . . . . . . .
5.6.2 Algorithm Parameters . . . . . . . . . . . . . . .
5.7 Optimal Box (Prototype) . . . . . . . . . . . . . . . . . .
5.7.1 Supported Domains . . . . . . . . . . . . . . . . .
5.7.2 Sample Class Parameters . . . . . . . . . . . . . .
5.7.3 Insert Membership Function Parameters . . . . . .
5.7.4 Feature Optimization Parameters . . . . . . . . . .
5.7.5 Optimization Settings Parameters . . . . . . . . .
5.7.6 Optimization Output Parameters . . . . . . . . . .
5.8 Classier . . . . . . . . . . . . . . . . . . . . . . . . . .
5.8.1 Supported Domains . . . . . . . . . . . . . . . . .
5.8.2 General Parameters . . . . . . . . . . . . . . . . .
5.8.3 Train: General Parameters . . . . . . . . . . . . .
5.8.4 Train: Feature Space . . . . . . . . . . . . . . . .
5.8.5 Train: Classier Parameters . . . . . . . . . . . .
5.8.6 Apply: General Parameters . . . . . . . . . . . . .
5.8.7 Apply: Feature Space . . . . . . . . . . . . . . . .
5.8.8 Query: General Parameters . . . . . . . . . . . . .
5.8.9 Query: Query Information about a trained classier

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

49
49
49
50
50
51
51
51
52
52
53
53
53
54
54
54
55
56
57
57
57
57
58
58
58
58
59
59
60
61
61
62
62
62
62
63
63
64
64

6 Template Matching
6.1 Overview Template Editor
6.2 Select Samples . . . . . .
6.3 Generate Templates . . . .
6.4 Test Template . . . . . . .
6.4.1 Subset Selection .
6.4.2 Template Selection
6.4.3 Test . . . . . . . .
6.4.4 Test Parameters . .
6.4.5 Review Targets . .
6.4.6 Template Quality .

.
.
.
.
.
.
.
.
.
.

66
66
67
68
69
70
70
70
70
71
72

.
.
.
.
.
.
.
.
.
.

Reference Book

11 June 2014

CONTENTS
6.5

Negative Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

7 Variables Operation Algorithm

7.1 Timer . . . . . . . . . . . . .
7.1.1 Supported Domains . .
7.1.2 Algorithm Parameters
7.2 Calculate Random Number . .
7.2.1 Algorithm Parameters
7.3 Update Variable . . . . . . . .
7.3.1 Supported Domains . .
7.3.2 Algorithm Parameters
7.4 Compute Statistical Value . . .
7.4.1 Supported Domains . .
7.4.2 Active Classes . . . .
7.4.3 Algorithm Parameters
7.5 Compose Text . . . . . . . . .
7.5.1 Supported Domains . .
7.5.2 Algorithm Parameters
7.6 Update Region . . . . . . . .
7.6.1 Supported Domains . .
7.6.2 Algorithm Parameters
7.7 Update Image Object List . . .
7.7.1 Supported Domains . .
7.7.2 Algorithm Parameters
7.8 Update Feature List . . . . . .
7.8.1 Supported Domains . .
7.9 Automatic Threshold . . . . .
7.9.1 Supported Domains . .
7.9.2 Algorithm Parameters
7.10 Update Array . . . . . . . . .
7.10.1 Supported Domains . .
7.10.2 Algorithm Parameters

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

73
73
73
73
73
74
74
74
74
76
76
76
76
77
77
77
78
78
78
80
80
80
80
81
81
82
82
83
83
83

8 Basic Object Reshaping Algorithms

8.1 Remove Objects . . . . . . . . .
8.1.1 Supported Domains . . .
8.1.2 Algorithm Parameters .
8.2 Merge Region . . . . . . . . . .
8.2.1 Supported Domains . . .
8.2.2 Algorithm Parameters .
8.3 Grow Region . . . . . . . . . .
8.3.1 Supported Domains . . .
8.3.2 Algorithm Parameters .
8.4 Convert to Sub-objects . . . . .
8.4.1 Supported Domains . . .
8.4.2 Algorithm Parameters .
8.5 Convert Image Objects . . . . .
8.5.1 Supported Domains . . .
8.5.2 Algorithm Parameters .
8.6 Cut Objects at Region . . . . . .
8.6.1 Supported Domains . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

84
84
84
84
85
85
86
86
86
87
87
87
87
87
87
88
89
89

Reference Book

11 June 2014

CONTENTS
8.6.2

vi
Algorithm Parameters . . . . . . . . . . . . . . . . . . . . . . . 89

9 Advanced Object Reshaping Algorithms

9.1 Shape Split (Prototype) . . . . . . . . . . . .
9.1.1 Supported Domains . . . . . . . . . .
9.1.2 Algorithm Parameters . . . . . . . .
9.2 Multiresolution Segmentation Region Grow .
9.2.1 Supported Domains . . . . . . . . . .
9.2.2 Image Layer Weights . . . . . . . . .
9.2.3 Thematic Layer Usage . . . . . . . .
9.2.4 Composition of Homogeneity Criteria
9.3 Image Object Fusion . . . . . . . . . . . . .
9.3.1 Supported Domains . . . . . . . . . .
9.3.2 Candidate Settings . . . . . . . . . .
9.3.3 Fitting Function . . . . . . . . . . . .
9.3.4 Weighted Sum . . . . . . . . . . . .
9.3.5 Merge Settings . . . . . . . . . . . .
9.3.6 Classication Settings . . . . . . . .
9.4 Border Optimization . . . . . . . . . . . . .
9.4.1 Supported Domains . . . . . . . . . .
9.4.2 Border Optimization Settings . . . .
9.4.3 Classication Settings . . . . . . . .
9.5 Morphology . . . . . . . . . . . . . . . . . .
9.5.1 Supported Domains . . . . . . . . . .
9.5.2 Morphology Settings . . . . . . . . .
9.5.3 Classication Settings . . . . . . . .
9.6 Watershed Transformation . . . . . . . . . .
9.6.1 Supported Domains . . . . . . . . . .
9.6.2 Watershed Settings . . . . . . . . . .
9.6.3 Classication Settings . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

91
91
91
91
92
92
92
92
92
93
93
93
94
95
95
96
96
96
97
97
98
98
98
99
100
100
100
100

10 Pixel-Based Object Reshaping Algorithms

10.1 Pixel-Based Object Resizing . . . . . . . . . . . . . .
10.1.1 Supported Domains . . . . . . . . . . . . . . .
10.1.2 Algorithm Parameters . . . . . . . . . . . . .
10.1.3 Growing/Shrinking Parameters . . . . . . . . .
10.1.4 Candidate Object Domain Parameters . . . . .
10.1.5 Pixel Level Constraint Parameters . . . . . . .
10.1.6 Candidate Surface Tension Parameters . . . . .
10.1.7 Size Limits Parameters . . . . . . . . . . . . .
10.2 Pixel-Based Density Filter . . . . . . . . . . . . . . .
10.2.1 Supported Domains . . . . . . . . . . . . . . .
10.2.2 Algorithm Parameters . . . . . . . . . . . . .
10.2.3 Growing and Shrinking Directions Parameters
10.2.4 Density Criteria Parameters . . . . . . . . . .
10.3 Pixel-Based Shape Processing Filters . . . . . . . . . .
10.3.1 Supported Domains . . . . . . . . . . . . . . .
10.3.2 Algorithm Parameters . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

102
102
102
102
104
104
104
106
109
109
109
110
110
111
113
113
113

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

11 Linking Operation Algorithms

120
11.1 Create Links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Reference Book

11 June 2014

CONTENTS

vii

11.1.1 Supported Domains . . . . . . . . . .

11.1.2 Algorithm Parameters . . . . . . . .
11.1.3 Candidate Object Domain Parameters
11.1.4 Overlap Settings Parameters . . . . .
11.2 Delete Links . . . . . . . . . . . . . . . . . .
11.2.1 Supported Domains . . . . . . . . . .
11.2.2 Algorithm Parameters . . . . . . . .

.
.
.
.
.
.
.

120
121
121
121
123
123
123

12 Level Operation Algorithms

12.1 Copy Image Object Level . . .
12.1.1 Supported Domains . .
12.1.2 Algorithm Parameters
12.2 Delete Image Object Level . .
12.2.1 Supported Domains . .
12.3 Rename Image Object Level .
12.3.1 Supported Domains . .
12.3.2 Algorithm Parameters

.
.
.
.
.
.
.
.

124
124
124
124
124
124
125
125
125

13 Map Operations Algorithms

13.1 Copy Map . . . . . . . . . . .
13.1.1 Supported Domains . .
13.1.2 Algorithm Parameters
13.2 Delete Map . . . . . . . . . .
13.2.1 Supported Domains . .
13.3 Synchronize Map . . . . . . .
13.3.1 Supported Domains . .
13.3.2 Algorithm Parameters
13.4 3D/4D Settings . . . . . . . .
13.4.1 Supported Domains . .
13.4.2 Algorithm Parameters
13.5 Scene Properties . . . . . . . .
13.5.1 Supported Domains . .
13.5.2 Algorithm Parameters
13.6 Rename Map . . . . . . . . .
13.6.1 Supported Domains . .
13.6.2 Algorithm Parameters
13.7 Transfer Layer . . . . . . . . .
13.7.1 Supported Domains . .
13.7.2 Algorithm Parameters

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

126
126
126
126
129
129
130
130
130
131
131
131
133
133
133
133
133
134
134
134
134

14 Image Layer Operation Algorithms

14.1 Distance Map . . . . . . . . . .
14.1.1 Supported Domains . . .
14.1.2 Algorithm Parameters .
14.2 Create Temporary Image Layer .
14.2.1 Supported Domains . . .
14.2.2 Algorithm Parameters .
14.3 Delete Layer . . . . . . . . . . .
14.3.1 Supported Domains . . .
14.3.2 Algorithm Parameters .
14.4 Convolution Filter . . . . . . . .

.
.
.
.
.
.
.
.
.
.

135
135
135
135
136
136
136
137
137
137
137

Reference Book

11 June 2014

CONTENTS
14.4.1 Supported Domains . . . . . . .
14.4.2 Algorithm Parameters . . . . .
14.4.3 Kernel Parameters . . . . . . .
14.4.4 Layers Parameters . . . . . . .
14.5 Layer Normalization . . . . . . . . . .
14.5.1 Supported Domains . . . . . . .
14.5.2 Algorithm Parameters . . . . .
14.5.3 Layers Parameters . . . . . . .
14.6 Median Filter . . . . . . . . . . . . . .
14.6.1 Supported Domains . . . . . . .
14.6.2 Kernel Parameters . . . . . . .
14.6.3 Layers Parameters . . . . . . .
14.7 Sobel Operation Filter . . . . . . . . . .
14.7.1 Supported Domains . . . . . . .
14.7.2 Kernel Parameters . . . . . . .
14.7.3 Layers Parameters . . . . . . .
14.8 Pixel Freq. Filter . . . . . . . . . . . .
14.8.1 Supported Domains . . . . . . .
14.8.2 Kernel Parameters . . . . . . .
14.8.3 Layers Parameters . . . . . . .
14.9 Pixel Min/Max Filter (Prototype) . . . .
14.9.1 Supported Domains . . . . . . .
14.9.2 Algorithm Parameters . . . . .
14.9.3 Kernel Parameters . . . . . . .
14.9.4 Layers Parameters . . . . . . .
14.10Edge Extraction Lee Sigma . . . . . . .
14.10.1 Supported Domains . . . . . . .
14.10.2 Algorithm Parameters . . . . .
14.11Edge Extraction Canny . . . . . . . . .
14.11.1 Supported Domains . . . . . . .
14.11.2 Algorithm Parameters . . . . .
14.12Edge 3D Filter . . . . . . . . . . . . . .
14.12.1 Supported Domains . . . . . . .
14.12.2 Algorithm Parameters . . . . .
14.12.3 Kernel Parameters . . . . . . .
14.12.4 Layer Parameters . . . . . . . .
14.13Surface Calculation . . . . . . . . . . .
14.13.1 Supported Domains . . . . . . .
14.13.2 Algorithm Parameters . . . . .
14.14Layer Arithmetics . . . . . . . . . . . .
14.14.1 Supported Domains . . . . . . .
14.14.2 Algorithm Parameters . . . . .
14.15Line Extraction . . . . . . . . . . . . .
14.15.1 Supported Domains . . . . . . .
14.15.2 Algorithm Parameters . . . . .
14.16Abs. Mean Deviation Filter (Prototype)
14.16.1 Supported Domains . . . . . . .
14.16.2 Algorithm Parameters . . . . .
14.16.3 Kernel Parameters . . . . . . .
14.16.4 Layers Parameters . . . . . . .
14.17Contrast Filter (Prototype) . . . . . . .

viii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

137
138
139
139
140
141
141
141
141
142
142
142
143
143
143
143
144
144
145
145
146
146
146
146
147
147
148
148
149
149
149
150
151
151
151
152
152
152
152
154
154
154
156
156
156
157
157
157
158
158
159

11 June 2014

CONTENTS

14.17.1 Supported Domains . . . . . . . . . . .

14.17.2 Algorithm Parameters . . . . . . . . .
14.18Pixel Filters Sliding Window . . . . . . . . . .
14.18.1 Supported Domains . . . . . . . . . . .
14.18.2 Algorithm Parameters . . . . . . . . .
14.18.3 Filter Kernel Sizes Selection Parameters
14.19Fill Pixel Values . . . . . . . . . . . . . . . . .
14.19.1 Supported Domains . . . . . . . . . . .
14.19.2 Algorithm Parameters . . . . . . . . .
14.20Template Matching . . . . . . . . . . . . . . .
14.20.1 Supported Domains . . . . . . . . . . .
14.20.2 Algorithm Parameters . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.

159
159
159
160
160
160
160
161
161
162
162
162

15 Point Cloud Algorithms

15.1 LiDAR File Converter . . . . . . . . . . .
15.1.1 Supported Domains . . . . . . . . .
15.1.2 Algorithm Parameters . . . . . . .
15.2 Merge point clouds . . . . . . . . . . . . .
15.2.1 Supported Domains . . . . . . . . .
15.2.2 Algorithm Parameters . . . . . . .
15.3 Create temporary point cloud . . . . . . . .
15.3.1 Supported Domains . . . . . . . . .
15.3.2 Algorithm Parameters . . . . . . .
15.4 Transfer temporary point cloud information
15.4.1 Supported Domains . . . . . . . . .
15.4.2 Algorithm Parameters . . . . . . .
15.5 Export point cloud . . . . . . . . . . . . .
15.5.1 Supported Domains . . . . . . . . .
15.5.2 Algorithm Parameters . . . . . . .
15.6 Export line from point cloud . . . . . . . .
15.6.1 Supported Domains . . . . . . . . .
15.6.2 Algorithm Parameters . . . . . . .
15.7 Cluster analysis of point cloud . . . . . . .
15.7.1 Supported Domains . . . . . . . . .
15.7.2 Algorithm Parameters . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

163
163
163
163
165
165
165
166
166
166
168
168
168
169
169
169
171
171
171
172
172
172

16 Thematic Layer Operation Algorithms

16.1 Create Temporary Thematic Vector Layer
16.1.1 Supported Domains . . . . . . . .
16.1.2 Algorithm Parameters . . . . . .
16.2 Assign Class by Thematic Layer . . . . .
16.2.1 Supported Domains . . . . . . . .
16.2.2 Algorithm Parameters . . . . . .
16.3 Synchronize Image Object Hierarchy . . .
16.3.1 Supported Domains . . . . . . . .
16.3.2 Algorithm Parameters . . . . . .
16.4 Read Thematic Attribute . . . . . . . . .
16.4.1 Supported Domains . . . . . . . .
16.4.2 Algorithm Parameters . . . . . .
16.5 Write Thematic Attributes . . . . . . . .
16.5.1 Supported Domains . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.

174
174
174
174
176
176
177
177
177
177
178
178
178
178
178

.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

11 June 2014

CONTENTS

16.5.2 Algorithm Parameters . . . . . .

16.6 Modify Thematic Attribute Column . . .
16.6.1 Supported Domains . . . . . . . .
16.6.2 Algorithm Parameters . . . . . .
16.7 Convert Image Objects to Vector Objects .
16.7.1 Supported Domains . . . . . . . .
16.7.2 Algorithm Parameters . . . . . .
16.8 Vector Boolean Operation . . . . . . . . .
16.8.1 Supported Domains . . . . . . . .
16.8.2 Algorithm Parameters . . . . . .
16.9 Vector Smoothing . . . . . . . . . . . . .
16.9.1 Supported Domains . . . . . . . .
16.9.2 Algorithm Parameters . . . . . .
16.9.3 Smooth mode . . . . . . . . . . .
16.10Vector Simplication . . . . . . . . . . .
16.10.1 Supported Domain . . . . . . . .
16.10.2 Algorithm Parameters . . . . . .
16.11Vector Orthogonalization . . . . . . . . .
16.11.1 Supported Domain . . . . . . . .
16.11.2 Algorithm Parameters . . . . . .
16.12Vector Integration . . . . . . . . . . . . .
16.12.1 Supported Domain . . . . . . . .
16.12.2 Algorithm Parameters . . . . . .
16.13Vector Remove Intersections . . . . . . .
16.13.1 Supported Domain . . . . . . . .
16.13.2 Algorithm Parameters . . . . . .
16.14Vector Dissolve . . . . . . . . . . . . . .
16.14.1 Supported Domain . . . . . . . .
16.14.2 Algorithm Parameters . . . . . .
16.15Delete Vector Objects . . . . . . . . . . .
16.15.1 Supported Domain . . . . . . . .
17 Workspace Automation Algorithms
17.1 Create Scene Copy . . . . . . . . .
17.1.1 Supported Domains . . . . .
17.1.2 Algorithm Parameters . . .
17.2 Create Scene Subset . . . . . . . . .
17.2.1 Supported Domains . . . . .
17.2.2 Algorithm Parameters . . .
17.3 Create Scene Tiles . . . . . . . . .
17.3.1 Supported Domains . . . . .
17.3.2 Tile Size Parameters . . . .
17.4 Submit Scenes for Analysis . . . . .
17.4.1 Supported Domains . . . . .
17.4.2 Algorithm Parameters . . .
17.4.3 Stitching Parameters . . . .
17.4.4 Post-Processing Parameters
17.5 Delete Scenes . . . . . . . . . . . .
17.5.1 Supported Domains . . . . .
17.5.2 Algorithm Parameters . . .
17.6 Read Subscene Statistics . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

178
179
179
179
180
180
180
180
180
181
181
181
182
182
182
182
182
183
183
183
184
184
184
185
185
185
185
185
185
185
186

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

187
187
187
187
188
188
188
190
190
190
190
190
190
192
192
193
193
193
193

Reference Book

11 June 2014

CONTENTS

17.6.1 Supported Domains . . . . . . . . . . . . . . . . . . . . . . . . . 194

17.6.2 Algorithm Parameters . . . . . . . . . . . . . . . . . . . . . . . 194
17.6.3 Mathematical Parameters . . . . . . . . . . . . . . . . . . . . . . 195
18 Interactive Operation Algorithms
18.1 Show User Warning . . . . . . . . .
18.1.1 Supported Domains . . . . .
18.1.2 Algorithm Parameters . . .
18.2 Set Active Pixel . . . . . . . . . . .
18.2.1 Supported Domains . . . . .
18.2.2 Algorithm Parameters . . .
18.3 Create/Modify Project . . . . . . .
18.3.1 Supported Domains . . . . .
18.3.2 Image Layer Parameters . .
18.3.3 Thematic Layer Parameters
18.3.4 General Settings Parameters
18.4 Manual Classication . . . . . . . .
18.4.1 Supported Domains . . . . .
18.4.2 Algorithm Parameters . . .
18.5 Congure Object Table . . . . . . .
18.5.1 Supported Domains . . . . .
18.5.2 Algorithm Parameters . . .
18.6 Select Input Mode . . . . . . . . . .
18.6.1 Supported Domains . . . . .
18.6.2 Algorithm Parameters . . .
18.7 Start Thematic Edit Mode . . . . . .
18.7.1 Supported Domains . . . . .
18.7.2 Algorithm Parameters . . .
18.8 Select Thematic Objects . . . . . .
18.8.1 Algorithm Parameters . . .
18.9 Finish Thematic Edit Mode . . . . .
18.9.1 Supported Domains . . . . .
18.9.2 Algorithm Parameters . . .
18.10Select Image Object . . . . . . . . .
18.10.1 Supported Domains . . . . .
18.10.2 Algorithm Parameters . . .
18.11Polygon Cut . . . . . . . . . . . . .
18.11.1 Supported Domain . . . . .
18.11.2 Algorithm Parameters . . .
18.12Save/Restore View Settings . . . . .
18.12.1 Supported Domain . . . . .
18.12.2 Algorithm Parameters . . .
18.13Display Map . . . . . . . . . . . . .
18.13.1 Supported Domains . . . . .
18.13.2 Algorithm Parameters . . .
18.14Dene View Layout . . . . . . . . .
18.14.1 Supported Domains . . . . .
18.14.2 Algorithm Parameters . . .
18.14.3 Pane Parameters . . . . . .
18.15Set Custom View Settings . . . . . .
18.15.1 Supported Domains . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

197
197
197
197
197
197
197
198
198
198
198
199
199
199
199
200
200
200
200
200
200
200
201
201
201
202
202
202
202
202
203
203
203
203
203
204
205
205
205
205
205
206
206
206
207
207
207

11 June 2014

CONTENTS
18.15.2 Algorithm Parameters
18.16Change Visible Layers . . . .
18.16.1 Supported Domains . .
18.16.2 Algorithm Parameters
18.17Change Visible Map . . . . .
18.17.1 Supported Domains . .
18.17.2 Algorithm Parameters
18.18Show Slide . . . . . . . . . .
18.18.1 Algorithm Parameters
18.19Ask Question . . . . . . . . .
18.19.1 Supported Domains . .
18.19.2 Algorithm Parameters
18.20Set Project State . . . . . . . .
18.20.1 Supported Domains . .
18.20.2 Algorithm Parameters
18.21Save/Restore Project State . .
18.21.1 Supported Domains . .
18.21.2 Algorithm Parameters
18.22Show HTML Help . . . . . .
18.22.1 Supported Domains . .
18.22.2 Algorithm Parameters
18.23Congure Image Equalization
18.23.1 Supported Domains . .
18.23.2 Algorithm Parameters
18.24Create/Update Class . . . . . .
18.24.1 Supported Domains . .
18.24.2 Algorithm Parameters

xii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

207
210
211
211
211
211
211
212
212
212
212
212
213
213
213
213
213
213
214
214
214
214
214
214
216
216
216

19 Parameter Set Operations Algorithms

19.1 Apply Parameter Set . . . . . . .
19.1.1 Supported Domains . . . .
19.1.2 Parameter Set Name . . .
19.2 Update Parameter Set . . . . . . .
19.2.1 Supported Domains . . . .
19.2.2 Parameter Set Name . . .
19.3 Load Parameter Set . . . . . . . .
19.3.1 Supported Domains . . . .
19.3.2 Algorithm Parameters . .
19.4 Save Parameter Set . . . . . . . .
19.4.1 Supported Domains . . . .
19.4.2 Algorithm Parameters . .
19.5 Delete Parameter Set File . . . . .
19.5.1 Supported Domains . . . .
19.5.2 Algorithm Parameters . .
19.6 Update Action from Parameter Set
19.6.1 Supported Domain . . . .
19.6.2 Algorithm Parameters . .
19.7 Update Parameter Set from Action
19.7.1 Supported Domains . . . .
19.7.2 Algorithm Parameters . .
19.8 Apply Active Action to Variables .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

218
218
218
218
218
218
219
219
219
219
219
220
220
220
220
220
220
221
221
221
221
221
221

Reference Book

11 June 2014

CONTENTS

xiii

19.8.1 Supported Domains . . . . . . . . . . . . . . . . . . . . . . . . . 221

19.8.2 Algorithm Parameters . . . . . . . . . . . . . . . . . . . . . . . 222
20 Sample Operation Algorithms
20.1 Classied Image Objects to Samples
20.1.1 Supported Domains . . . . .
20.1.2 Algorithm Parameters . . .
20.2 Cleanup Redundant Samples . . . .
20.2.1 Supported Domains . . . . .
20.2.2 Algorithm Parameters . . .
20.3 Nearest Neighbor Conguration . .
20.3.1 Supported Domains . . . . .
20.3.2 Algorithm Parameters . . .
20.4 Delete All Samples . . . . . . . . .
20.4.1 Supported Domains . . . . .
20.5 Delete Samples of Classes . . . . .
20.5.1 Supported Domains . . . . .
20.5.2 Algorithm Parameters . . .
20.6 Disconnect All Samples . . . . . . .
20.6.1 Supported Domains . . . . .
20.7 Sample Selection . . . . . . . . . .
20.7.1 Supported Domains . . . . .
20.7.2 Algorithm Parameters . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

223
223
223
223
223
223
223
224
224
224
224
224
224
224
224
225
225
225
225
225

21 Export Algorithms
21.1 Export Classication View . . . .
21.1.1 Supported Domains . . . .
21.1.2 Algorithm Parameters . .
21.2 Export Current View . . . . . . .
21.2.1 Supported Domains . . . .
21.2.2 Algorithm Parameters . .
21.2.3 Slices Parameters . . . . .
21.2.4 Frames Parameters . . . .
21.3 Export Thematic Raster Files . . .
21.3.1 Supported Domains . . . .
21.3.2 Algorithm Parameters . .
21.4 Export Existing Vector Layer . . .
21.4.1 Supported Domains . . . .
21.4.2 Algorithm Parameters . .
21.5 Export Domain Statistics . . . . .
21.5.1 Supported Domains . . . .
21.5.2 Algorithm Parameters . .
21.5.3 Statistical Operations . . .
21.6 Export Project Statistics . . . . . .
21.6.1 Supported Domains . . . .
21.6.2 Algorithm Parameters . .
21.7 Export Object Statistics . . . . . .
21.7.1 Supported Domains . . . .
21.7.2 Algorithm Parameters . .
21.7.3 Report Parameters . . . .
21.8 Export Object Statistics for Report

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

226
226
226
226
227
227
227
228
229
229
229
229
230
230
231
231
231
232
232
233
233
233
233
234
234
235
235

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

11 June 2014

CONTENTS

xiv

21.8.1 Supported Domains . . . .

21.8.2 Algorithm Parameters . .
21.9 Export Vector Layer . . . . . . . .
21.9.1 Supported Domains . . . .
21.9.2 Algorithm Parameters . .
21.9.3 Export Data Parameters .
21.9.4 Export Format Parameters
21.10Export Image Object View . . . .
21.10.1 Supported Domains . . . .
21.10.2 Output Parameters . . . .
21.10.3 Settings . . . . . . . . . .
21.11Export Mask Image . . . . . . . .
21.11.1 Supported Domains . . . .
21.11.2 Algorithm Parameters . .
21.12Export Image . . . . . . . . . . .
21.12.1 Supported Domains . . . .
21.12.2 Algorithm Parameters . .
21.13Export Result Preview . . . . . .
21.13.1 Supported Domains . . . .
21.13.2 Algorithm Parameters . .
22 Image Registration Algorithms
22.1 Image Registration . . . . . .
22.1.1 Supported Domains . .
22.1.2 Algorithm Parameters
22.2 Delete Landmarks . . . . . . .
22.2.1 Supported Domains . .
22.3 Set Landmark . . . . . . . . .
22.3.1 Supported Domains . .

.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

235
235
236
236
236
236
237
237
237
238
238
239
239
239
240
240
240
241
242
242

.
.
.
.
.
.
.

243
243
244
244
245
245
245
245

23 About Features
246
23.1 About Features as a Source of Information . . . . . . . . . . . . . . . . . 246
23.1.1 Conversions of Feature Values . . . . . . . . . . . . . . . . . . . 246
23.2 Object Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
24 Vector Features
248
24.1 Vector object attribute . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
24.1.1 Editable Parameters . . . . . . . . . . . . . . . . . . . . . . . . 248
25 Object Features: Customized
25.1 Create Customized Features . . . . . . . . . .
25.2 Arithmetic Customized Features . . . . . . . .
25.3 Relational Customized Features . . . . . . . .
25.3.1 Relations Between Surrounding Objects
25.3.2 Relational Functions . . . . . . . . . .
25.4 Finding Customized Features . . . . . . . . . .

.
.
.
.
.
.

249
249
249
251
252
253
255

26 Object Features: Type

256
26.1 Is 3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
26.2 Is Connected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
27 Object Features: Layer Values

257

Reference Book

11 June 2014

CONTENTS
27.1 Mean . . . . . . . . . . . . . . . . . . .
27.1.1 Brightness . . . . . . . . . . . .
27.1.2 Layer 1/2/3 . . . . . . . . . . . .
27.1.3 Max. Diff. . . . . . . . . . . . .
27.2 Quantile . . . . . . . . . . . . . . . . . .
27.2.1 Editable Parameters . . . . . . .
27.2.2 Parameters . . . . . . . . . . . .
27.2.3 Feature Value Range . . . . . . .
27.3 Mode . . . . . . . . . . . . . . . . . . .
27.3.1 Editable Parameters . . . . . . .
27.4 Standard Deviation . . . . . . . . . . . .
27.4.1 Layer 1/2/3 . . . . . . . . . . . .
27.5 Skewness . . . . . . . . . . . . . . . . .
27.5.1 Layer Values . . . . . . . . . . .
27.6 Pixel Based . . . . . . . . . . . . . . . .
27.6.1 Ratio . . . . . . . . . . . . . . .
27.6.2 Min. Pixel Value . . . . . . . . .
27.6.3 Max. Pixel Value . . . . . . . . .
27.6.4 Mean of Inner Border . . . . . . .
27.6.5 Mean of Outer Border . . . . . .
27.6.6 Border Contrast . . . . . . . . . .
27.6.7 Contrast to Neighbor Pixels . . .
27.6.8 Edge Contrast of Neighbor Pixels
27.6.9 StdDev. to Neighbor Pixels . . . .
27.6.10 Circular Mean . . . . . . . . . .
27.6.11 Circular StdDev . . . . . . . . .
27.6.12 Circular Std Dev/Mean . . . . . .
27.7 To Neighbors . . . . . . . . . . . . . . .
27.7.1 Mean Diff. to Neighbors . . . . .
27.7.2 Mean Diff. to Neighbors (Abs) . .
27.7.3 Mean Diff. to Darker Neighbors .
27.7.4 Mean Diff. to Brighter Neighbors
27.7.5 Number of Brighter Objects . . .
27.7.6 Number of Darker Objects . . . .
27.7.7 Rel. Border to Brighter Neighbors
27.8 To Superobject . . . . . . . . . . . . . .
27.8.1 Mean Diff. to Superobject . . . .
27.8.2 Ratio to Superobject . . . . . . .
27.8.3 Std. Dev. Diff. to Superobject . .
27.8.4 Std. Dev. Ratio to Superobject . .
27.9 To Scene . . . . . . . . . . . . . . . . . .
27.9.1 Mean Diff. to Scene . . . . . . .
27.9.2 Ratio To Scene . . . . . . . . . .
27.10Hue, Saturation, Intensity . . . . . . . . .
27.10.1 HSI Transformation . . . . . . .

xv
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

257
257
258
259
259
260
260
260
260
260
261
261
261
262
262
262
263
264
265
266
266
267
268
270
270
271
271
272
272
273
274
275
276
276
276
277
277
277
278
279
279
279
280
281
281

28 Object Features: Geometry

28.1 Extent . . . . . . . . . . . . . . . . . . . . . .
28.1.1 Area . . . . . . . . . . . . . . . . . . .
28.1.2 Border Length [for 2D Image Objects] .
28.1.3 Border Length [for 3D Image Objects] .

.
.
.
.

282
282
282
283
284

Reference Book

11 June 2014

CONTENTS

xvi

28.1.4 Length [for 2D Image Objects] . . . . . . . . . . . . . . . .

28.1.5 Length [for 3D Image Objects] . . . . . . . . . . . . . . . .
28.1.6 Length/Thickness . . . . . . . . . . . . . . . . . . . . . . .
28.1.7 Length/Width [for 2D Image Objects] . . . . . . . . . . . .
28.1.8 Length/Width [for 3D Image Objects] . . . . . . . . . . . .
28.1.9 Number of Pixels . . . . . . . . . . . . . . . . . . . . . . .
28.1.10 Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . .
28.1.11 Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28.1.12 Width [for 2D Image Objects] . . . . . . . . . . . . . . . .
28.1.13 Width [for 3D Image Objects] . . . . . . . . . . . . . . . .
28.2 Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28.2.1 Asymmetry [for 2D Image Objects] . . . . . . . . . . . . .
28.2.2 Asymmetry [for 3D Image Objects] . . . . . . . . . . . . .
28.2.3 Border Index . . . . . . . . . . . . . . . . . . . . . . . . .
28.2.4 Compactness [for 2D Image Objects] . . . . . . . . . . . .
28.2.5 Compactness [for 3D Image Objects] . . . . . . . . . . . .
28.2.6 Density [for 2D Image Objects] . . . . . . . . . . . . . . .
28.2.7 Density [for 3D Image Objects] . . . . . . . . . . . . . . .
28.2.8 Elliptic Fit [for 2D Image Objects] . . . . . . . . . . . . . .
28.2.9 Elliptic Fit [for 3D Image Objects] . . . . . . . . . . . . . .
28.2.10 Main Direction [for 2D Image Objects] . . . . . . . . . . .
28.2.11 Main Direction [for 3D Image Objects] . . . . . . . . . . .
28.2.12 Radius of Largest Enclosed Ellipse [for 2D Image Objects] .
28.2.13 Radius of Largest Enclosed Ellipse [for 3D Image Objects] .
28.2.14 Radius of Smallest Enclosing Ellipse [for 2D Image Objects]
28.2.15 Radius of Smallest Enclosing Ellipse [for 3D Image Objects]
28.2.16 Rectangular Fit [for 2D Image Objects] . . . . . . . . . . .
28.2.17 Rectangular Fit [for 3D Image Objects] . . . . . . . . . . .
28.2.18 Roundness [for 2D Image Objects] . . . . . . . . . . . . . .
28.2.19 Roundness [for 3D Image Objects] . . . . . . . . . . . . . .
28.2.20 Shape Index [for 2D Image Objects] . . . . . . . . . . . . .
28.2.21 Shape Index [for 3D Image Objects] . . . . . . . . . . . . .
28.3 To Super-Object . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28.3.1 Rel. Area to Super-Object . . . . . . . . . . . . . . . . . .
28.3.2 Rel. Rad. Position to Super-Object . . . . . . . . . . . . . .
28.3.3 Rel. Inner Border to Super-Object . . . . . . . . . . . . . .
28.3.4 Distance to Super-Object Center . . . . . . . . . . . . . . .
28.3.5 Elliptic Distance to Super-Object Center . . . . . . . . . . .
28.3.6 Is End of Super-Object . . . . . . . . . . . . . . . . . . . .
28.3.7 Is Center of Super-Object . . . . . . . . . . . . . . . . . .
28.3.8 Rel. X Position to Super-Object . . . . . . . . . . . . . . .
28.3.9 Rel. Y Position to Super-Object . . . . . . . . . . . . . . .
28.4 Based on Polygons . . . . . . . . . . . . . . . . . . . . . . . . . .
28.4.1 Edges Longer Than . . . . . . . . . . . . . . . . . . . . . .
28.4.2 Number of Right Angles With Edges Longer Than . . . . .
28.4.3 Area (Excluding Inner Polygons) . . . . . . . . . . . . . .
28.4.4 Area (Including Inner Polygons) . . . . . . . . . . . . . . .
28.4.5 Average Length of Edges (Polygon) . . . . . . . . . . . . .
28.4.6 Compactness (Polygon) . . . . . . . . . . . . . . . . . . .
28.4.7 Length of Longest Edge (Polygon) . . . . . . . . . . . . . .
28.4.8 Number of Edges (Polygon) . . . . . . . . . . . . . . . . .

Reference Book

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

284
285
285
285
286
287
287
287
288
288
288
289
289
290
291
292
292
293
293
294
295
296
296
297
298
299
299
300
301
302
302
303
303
303
304
305
306
307
307
308
308
308
309
309
309
310
311
311
311
312
312

11 June 2014

CONTENTS

xvii

28.4.9 Number of Inner Objects (Polygon) . . .

28.4.10 Perimeter (Polygon) . . . . . . . . . . .
28.4.11 Polygon Self-Intersection (Polygon) . . .
28.4.12 Std. Dev. of Length of Edges (Polygon) .
28.5 Based on Skeletons . . . . . . . . . . . . . . . .
28.5.1 Number of Segments of Order . . . . . .
28.5.2 Number of Branches of Order . . . . . .
28.5.3 Average Length of Branches of Order . .
28.5.4 Number of Branches of Length . . . . . .
28.5.5 Average Branch Length . . . . . . . . .
28.5.6 Avrg. Area Represented by Segments . .
28.5.7 Curvature/Length (Only Main Line) . . .
28.5.8 Degree of Skeleton Branching . . . . . .
28.5.9 Length of Main Line (No Cycles) . . . .
28.5.10 Length of Main Line (Regarding Cycles)
28.5.11 Length/Width (Only Main Line) . . . . .
28.5.12 Maximum Branch Length . . . . . . . .
28.5.13 Number of Segments . . . . . . . . . . .
28.5.14 Stddev Curvature (Only Main Line) . . .
28.5.15 Stddev of Area Represented by Segments
28.5.16 Width (Only Main Line) . . . . . . . . .
29 Object Features: Position
29.1 Distance . . . . . . . . . . . . . . . . . . .
29.1.1 Distance to Line . . . . . . . . . .
29.1.2 Distance to Scene Border . . . . . .
29.1.3 T Distance to First Frame (Pxl) . .
29.1.4 T Distance to Last Frame . . . . . .
29.1.5 X Distance to Scene Left Border . .
29.1.6 X Distance to Scene Right Border .
29.1.7 Y Distance to Scene Bottom Border
29.1.8 Y Distance to Scene Top Border . .
29.1.9 Z Distance to First Slice (Pxl) . . .
29.1.10 Z Distance to Last Slice (Pxl) . . .
29.2 Coordinate . . . . . . . . . . . . . . . . . .
29.2.1 Is at Active Pixel . . . . . . . . . .
29.2.2 Time (Pxl) . . . . . . . . . . . . .
29.2.3 Time Max (Pxl) . . . . . . . . . . .
29.2.4 Time Min (Pxl) . . . . . . . . . . .
29.2.5 X Center . . . . . . . . . . . . . .
29.2.6 X Max. . . . . . . . . . . . . . . .
29.2.7 X Min. . . . . . . . . . . . . . . .
29.2.8 Y Center . . . . . . . . . . . . . .
29.2.9 Y Max. . . . . . . . . . . . . . . .
29.2.10 Y Min. . . . . . . . . . . . . . . .
29.2.11 Z Center . . . . . . . . . . . . . .
29.2.12 Z Max . . . . . . . . . . . . . . . .
29.2.13 Z Min . . . . . . . . . . . . . . . .
29.3 Is Object in Region . . . . . . . . . . . . .
29.3.1 Editable Parameters . . . . . . . .
29.3.2 Feature Value Range . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

312
312
312
313
314
314
314
315
315
315
316
316
316
317
317
317
317
318
318
318
318

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

320
320
320
321
322
322
322
323
323
324
325
325
325
325
325
326
326
327
327
328
328
329
330
331
331
332
332
332
332

Reference Book

11 June 2014

CONTENTS

xviii

30 Object Features: Texture

30.1 Layer Value Texture Based on Sub-objects . . . . . . .
30.1.1 Mean of Sub-objects: Std. Dev. . . . . . . . .
30.1.2 Avrg. Mean Diff. to Neighbors of Sub-objects
30.2 Shape Texture Based on Sub-objects . . . . . . . . . .
30.2.1 Area of Sub-objects: Mean . . . . . . . . . . .
30.2.2 Area of Sub-objects: Std. Dev. . . . . . . . . .
30.2.3 Density of Sub-objects: Mean . . . . . . . . .
30.2.4 Density of Sub-objects: Std. Dev. . . . . . . .
30.2.5 Asymmetry of Sub-objects: Mean . . . . . . .
30.2.6 Asymmetry of Sub-objects: Std. Dev. . . . . .
30.2.7 Direction of Sub-objects: Mean . . . . . . . .
30.2.8 Direction of Sub-objects: Std. Dev. . . . . . .
30.3 Texture After Haralick . . . . . . . . . . . . . . . . .
30.3.1 Calculation of GLCM . . . . . . . . . . . . .
30.3.2 Parameters . . . . . . . . . . . . . . . . . . .
30.3.3 Expression . . . . . . . . . . . . . . . . . . .
30.3.4 References . . . . . . . . . . . . . . . . . . .
30.3.5 GLCM Homogeneity . . . . . . . . . . . . . .
30.3.6 GLCM Contrast . . . . . . . . . . . . . . . .
30.3.7 GLCM Dissimilarity . . . . . . . . . . . . . .
30.3.8 GLCM Entropy . . . . . . . . . . . . . . . . .
30.3.9 GLCM Ang. 2nd Moment . . . . . . . . . . .
30.3.10 GLCM Mean . . . . . . . . . . . . . . . . . .
30.3.11 GLCM Std. Dev. . . . . . . . . . . . . . . . .
30.3.12 GLCM Correlation . . . . . . . . . . . . . . .
30.3.13 GLDV Angular 2nd Moment . . . . . . . . . .
30.3.14 GLDV Entropy . . . . . . . . . . . . . . . . .
30.3.15 GLDV Mean . . . . . . . . . . . . . . . . . .
30.3.16 GLDV Contrast . . . . . . . . . . . . . . . . .
30.3.17 GLCM/GLDV . . . (Quick 8/11) . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

333
333
333
334
335
335
335
336
337
337
338
338
339
339
341
341
342
342
342
343
343
344
344
345
345
346
346
347
347
348
348

31 Object Features: Variables

349
31.1 Object Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
31.1.1 Editable Parameters . . . . . . . . . . . . . . . . . . . . . . . . 349
32 Object Features: Hierarchy
32.1 Level . . . . . . . . . . . .
32.2 Level Number . . . . . . . .
32.2.1 Parameters . . . . .
32.2.2 Expression . . . . .
32.2.3 Feature Value Range
32.2.4 Conditions . . . . .
32.3 Number of Higher Levels . .
32.3.1 Parameters . . . . .
32.3.2 Expression . . . . .
32.3.3 Feature Value Range
32.4 Number of Neighbors . . . .
32.4.1 Parameters . . . . .
32.4.2 Expression . . . . .
32.4.3 Feature Value Range

.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

.
.
.
.
.
.
.
.
.
.
.
.
.
.

350
350
350
350
350
350
350
351
351
351
351
351
351
351
351

11 June 2014

CONTENTS
32.5 Number of Sub-Objects . . .
32.5.1 Parameters . . . . .
32.5.2 Expression . . . . .
32.5.3 Feature Value Range
32.6 Number of Sublevels . . . .
32.6.1 Parameters . . . . .
32.6.2 Expression . . . . .
32.6.3 Feature Value Range

xix
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

351
352
352
352
352
352
352
352

33 Object Features: Thematic Attributes

33.1 Minimum overlap [%] with thematic polygons .
33.1.1 Editable Parameter . . . . . . . . . . .
33.1.2 Feature Value Range . . . . . . . . . .
33.2 Maximum overlap [%] with thematic polygons
33.2.1 Editable Parameter . . . . . . . . . . .
33.2.2 Feature Value Range . . . . . . . . . .
33.3 Number of Overlapping Thematic Objects . . .
33.3.1 Editable Parameter . . . . . . . . . . .
33.3.2 Feature Value Range . . . . . . . . . .
33.4 Thematic Object Attribute . . . . . . . . . . .
33.4.1 Editable Parameters . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.

353
353
353
353
353
354
354
354
354
354
354
354

34 Object Features: Point Cloud Features

34.1 Intensity . . . . . . . . . . . . . . .
34.1.1 Parameters . . . . . . . . .
34.2 X coordinate . . . . . . . . . . . . .
34.2.1 Parameters . . . . . . . . .
34.3 Y coordinate . . . . . . . . . . . . .
34.3.1 Parameters . . . . . . . . .
34.4 Elevation . . . . . . . . . . . . . .
34.4.1 Parameters . . . . . . . . .
34.5 Return Number . . . . . . . . . . .
34.5.1 Parameters . . . . . . . . .
34.6 Number of Returns . . . . . . . . .
34.6.1 Parameters . . . . . . . . .
34.7 Red . . . . . . . . . . . . . . . . .
34.7.1 Parameters . . . . . . . . .
34.8 Green . . . . . . . . . . . . . . . .
34.8.1 Parameters . . . . . . . . .
34.9 Blue . . . . . . . . . . . . . . . . .
34.9.1 Parameters . . . . . . . . .
34.10Number of points . . . . . . . . . .
34.10.1 Parameters . . . . . . . . .
34.11Class . . . . . . . . . . . . . . . . .
34.11.1 Parameters . . . . . . . . .
34.12Distance to Camera . . . . . . . . .
34.12.1 Parameters . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

355
355
355
357
357
358
358
360
360
361
361
363
363
364
364
366
366
367
367
369
369
371
371
372
372

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

35 Class-Related Features
375
35.1 Relations to Neighbor Objects . . . . . . . . . . . . . . . . . . . . . . . 375
35.1.1 Existence of . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

Reference Book

11 June 2014

CONTENTS

35.1.2 Number Of . . . . . . .
35.1.3 Border To . . . . . . . .
35.1.4 Rel. Border to . . . . .
35.1.5 Rel. Area of . . . . . . .
35.1.6 Distance to . . . . . . .
35.1.7 Mean Diff. to . . . . . .
35.1.8 Overlap of two Objects .
35.2 Relations to Sub-Objects . . . .
35.2.1 Existence Of . . . . . .
35.2.2 Number of . . . . . . .
35.2.3 Area of . . . . . . . . .
35.2.4 Rel. Area of . . . . . . .
35.2.5 Clark Aggregation Index
35.3 Relations to Superobjects . . . .
35.3.1 Existence of . . . . . . .
35.4 Relations to Classication . . .
35.4.1 Membership to . . . . .
35.4.2 Classied as . . . . . .
35.4.3 Classication Value of .
35.4.4 Class Name . . . . . . .
35.4.5 Class Color . . . . . . .
35.4.6 Assigned Class . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

376
376
377
378
379
379
380
381
381
381
382
382
383
383
384
384
384
385
385
386
386
386

.
.
.
.
.
.

388
388
388
388
388
389
390

37 Scene Features
37.1 Scene Variables . . . . . . . . . . . . . . . .
37.1.1 Editable Parameters . . . . . . . . .
37.2 Class-Related . . . . . . . . . . . . . . . . .
37.2.1 Number of Classied Objects . . . .
37.2.2 Number of Samples Per Class . . . .
37.2.3 Area of Classied Objects . . . . . .
37.2.4 Layer Mean of Classied Objects . .
37.2.5 Layer Std. Dev. of Classied Objects
37.2.6 Statistic of Object Value . . . . . . .
37.2.7 Class Variables . . . . . . . . . . . .
37.3 Scene-Related . . . . . . . . . . . . . . . . .
37.3.1 Existence of Object Level . . . . . .
37.3.2 Existence of Image Layer . . . . . .
37.3.3 Existence of Thematic Layer . . . . .
37.3.4 Existence of Map . . . . . . . . . . .
37.3.5 Mean of Scene . . . . . . . . . . . .
37.3.6 Standard deviation . . . . . . . . . .
37.3.7 Smallest Actual Pixel Value . . . . .
37.3.8 Largest Actual Pixel Value . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

391
391
391
391
392
392
392
393
393
394
394
395
395
395
395
396
396
397
397
397

36 Linked Object Features

36.1 Linked Object Count . . . .
36.1.1 Editable Parameters
36.2 Linked Objects Statistics . .
36.2.1 Editable Parameters
36.3 Linked Weight to PPO . . .
36.3.1 Editable Parameters

.
.
.
.
.
.

Reference Book

11 June 2014

CONTENTS

xxi

37.3.9 Validity of Region . . . . . . . . . . . . . .

37.3.10 Mode . . . . . . . . . . . . . . . . . . . . .
37.3.11 Quantile . . . . . . . . . . . . . . . . . . . .
37.3.12 Active Pixel t-value . . . . . . . . . . . . . .
37.3.13 Active Pixel x-value . . . . . . . . . . . . .
37.3.14 Active Pixel y-value . . . . . . . . . . . . .
37.3.15 Active Pixel z-value . . . . . . . . . . . . .
37.3.16 Image Directory . . . . . . . . . . . . . . .
37.3.17 Image Full Path . . . . . . . . . . . . . . . .
37.3.18 Image Name . . . . . . . . . . . . . . . . .
37.3.19 Is Active Vector Layer Changed . . . . . . .
37.3.20 Map Origin T . . . . . . . . . . . . . . . . .
37.3.21 Map Origin X . . . . . . . . . . . . . . . . .
37.3.22 Map Origin Y . . . . . . . . . . . . . . . . .
37.3.23 Map Origin Z . . . . . . . . . . . . . . . . .
37.3.24 Map Size T . . . . . . . . . . . . . . . . . .
37.3.25 Map Size X . . . . . . . . . . . . . . . . . .
37.3.26 Map Size Y . . . . . . . . . . . . . . . . . .
37.3.27 Map Size Z . . . . . . . . . . . . . . . . . .
37.3.28 Number of Image Layers . . . . . . . . . . .
37.3.29 Number of Maps . . . . . . . . . . . . . . .
37.3.30 Number of Objects . . . . . . . . . . . . . .
37.3.31 Number of Pixels in Scene . . . . . . . . . .
37.3.32 Number of Samples . . . . . . . . . . . . . .
37.3.33 Number of Thematic Layers . . . . . . . . .
37.3.34 Original Scene ID . . . . . . . . . . . . . .
37.3.35 Original Scene Name . . . . . . . . . . . . .
37.3.36 Scene ID . . . . . . . . . . . . . . . . . . .
37.3.37 Scene Magnication . . . . . . . . . . . . .
37.3.38 Scene Name . . . . . . . . . . . . . . . . .
37.3.39 Scene Pixel Size . . . . . . . . . . . . . . .
37.3.40 Scene Resolution . . . . . . . . . . . . . . .
37.3.41 Scene Bit Depth . . . . . . . . . . . . . . .
37.3.42 Second Level Scene Name . . . . . . . . . .
37.3.43 Slice Distance . . . . . . . . . . . . . . . . .
37.3.44 Time Series Distance . . . . . . . . . . . . .
37.3.45 Top Scene ID . . . . . . . . . . . . . . . . .
37.3.46 Top Scene Name . . . . . . . . . . . . . . .
37.3.47 User Name . . . . . . . . . . . . . . . . . .
37.3.48 Workspace Full Path . . . . . . . . . . . . .
37.4 Rule-Set Related . . . . . . . . . . . . . . . . . . .
37.4.1 Rule set items . . . . . . . . . . . . . . . . .
37.4.2 Class Variable . . . . . . . . . . . . . . . . .
37.4.3 Level Variable . . . . . . . . . . . . . . . .
37.4.4 Map Name Variable . . . . . . . . . . . . .
37.4.5 Image Layer Variable . . . . . . . . . . . . .
37.4.6 Thematic Layer Variable . . . . . . . . . . .
37.4.7 Number of Features in Feature List . . . . .
37.4.8 Number of Specied Features in Feature List
37.4.9 Rule Set Array Values . . . . . . . . . . . .
37.4.10 Rule Set Array Size . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Reference Book

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

398
398
398
399
399
399
400
400
400
400
400
400
401
401
401
401
401
402
402
403
403
403
403
404
404
404
404
405
405
405
405
405
405
405
406
406
406
406
406
406
407
407
407
407
407
407
407
408
408
408
408

11 June 2014

CONTENTS

xxii

37.4.11 Rule Set Array Item

37.5 Architect Related . . . . . .
37.5.1 Number of Actions .
37.6 File System . . . . . . . . .
37.6.1 Resolve Path . . . .
37.7 UI Related . . . . . . . . . .
37.7.1 Equalization . . . .

.
.
.
.
.
.
.

408
408
408
409
409
409
410

38 Process Related Features

38.1 Customized . . . . . . . . .
38.1.1 Diff. PPO . . . . . .
38.1.2 Ratio PPO . . . . .
38.2 Border to PPO . . . . . . . .
38.2.1 Editable Parameters
38.2.2 Parameters . . . . .
38.2.3 Expression . . . . .
38.2.4 Feature Value Range
38.3 Distance to PPO . . . . . . .
38.3.1 Editable Parameter .
38.4 Elliptic Dist. from PPO . . .
38.4.1 Parameters . . . . .
38.4.2 Editable Parameters
38.4.3 Expression . . . . .
38.4.4 Feature Value Range
38.5 Rel. border to PPO . . . . .
38.5.1 Editable Parameters
38.5.2 Parameters . . . . .
38.5.3 Expression . . . . .
38.5.4 Feature Value Range
38.6 Same Superobject as PPO . .
38.6.1 Editable Parameters
38.6.2 Parameters . . . . .
38.6.3 Expression . . . . .
38.6.4 Feature Value Range
38.7 Series ID . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

411
411
411
412
413
413
413
413
413
413
414
414
414
414
414
414
414
415
415
415
415
415
415
415
415
416
416

39 Region Features
39.1 Region-Related . . . . . . . . . . .
39.1.1 Number of Pixels in Region
39.1.2 T Extent . . . . . . . . . . .
39.1.3 T Origin . . . . . . . . . . .
39.1.4 X Extent . . . . . . . . . .
39.1.5 X Origin . . . . . . . . . .
39.1.6 Y Extent . . . . . . . . . .
39.1.7 Y Origin . . . . . . . . . .
39.1.8 Z Extent . . . . . . . . . . .
39.1.9 Z Origin . . . . . . . . . . .
39.2 Layer-Related . . . . . . . . . . . .
39.2.1 Mean . . . . . . . . . . . .
39.2.2 Standard Deviation . . . . .
39.3 Class-Related . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.

417
417
417
418
418
418
419
419
419
420
420
421
421
421
422

Reference Book

11 June 2014

CONTENTS

xxiii

39.3.1 Area of Classied Objects . . . . . . . . . . . . . . . . . . . . . 422

39.3.2 Relative Area of Classied Objects . . . . . . . . . . . . . . . . 423
40 Image Registration Features
40.1 Object-Related . . . . . . . . . . . .
40.1.1 Object Landmarks on the Map
40.2 Scene-Related . . . . . . . . . . . . .
40.2.1 Landmarks on the Map . . . .

.
.
.
.

424
424
424
424
424

41 Metadata
41.1 Active Slice Metadata item .
41.1.1 Editable Parameters
41.2 Image Layer Metadata item .
41.2.1 Editable Parameters
41.3 Metadata Item . . . . . . . .
41.3.1 Editable Parameters

.
.
.
.
.
.

425
425
425
425
425
426
426

.
.
.
.
.
.

42 Feature Variables
427
42.1 Feature Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
42.1.1 Editable Parameters . . . . . . . . . . . . . . . . . . . . . . . . 427
43 Widget Parameters for Architect Action Libraries
43.1 Add Checkbox . . . . . . . . . . . . . . . . .
43.2 Add Drop-down List . . . . . . . . . . . . . .
43.3 Add Button . . . . . . . . . . . . . . . . . . .
43.4 Add Radio Button Row . . . . . . . . . . . . .
43.5 Add Toolbar . . . . . . . . . . . . . . . . . . .
43.6 Add Editbox . . . . . . . . . . . . . . . . . . .
43.7 Add Editbox With Slider . . . . . . . . . . . .
43.8 Add Select Class . . . . . . . . . . . . . . . .
43.9 Add Select Feature . . . . . . . . . . . . . . .
43.10Add Select Multiple Features . . . . . . . . . .
43.11Add Select File . . . . . . . . . . . . . . . . .
43.12Add Select Level . . . . . . . . . . . . . . . .
43.13Add Select Image Layer . . . . . . . . . . . .
43.14Add Select Thematic Layer . . . . . . . . . . .
43.15Add Select Folder . . . . . . . . . . . . . . . .
43.16Add Slider . . . . . . . . . . . . . . . . . . . .
43.17Add Edit Layer Names . . . . . . . . . . . . .
43.18Add Layer Drop-down List . . . . . . . . . . .
43.19Add Manual Classication Buttons . . . . . . .
43.20Add Select Array Items . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

428
428
428
429
429
430
430
431
431
432
432
433
433
434
434
435
435
436
436
436
437

44 General Reference
44.1 Use Variables as Features . . . . . . . . . . . . . . . . .
44.2 About Metadata as a Source of Information . . . . . . .
44.2.1 Convert Metadata and Add it to the Feature Tree
44.3 General Reference . . . . . . . . . . . . . . . . . . . .
44.3.1 Rendering a Displayed Image . . . . . . . . . .

.
.
.
.
.

438
438
438
438
439
439

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Acknowledgments
443
The Visualization Toolkit (VTK) Copyright . . . . . . . . . . . . . . . . . . . 443

Reference Book

11 June 2014

CONTENTS
ITK Copyright . . . . . . . . . . .
python/tests/test_doctests.py
src/Verson.rc . . . . . . . .
src/gt_wkt_srs.cpp . . . . .

xxiv
.
.
.
.

.
.
.
.

Reference Book

.
.
.
.

444
444
445
445

11 June 2014

1 Introduction
eCognition Documentation is divided into two different books: the User Guide and the
Reference Book.
This Reference Book is a list of algorithms and features where you can look up how
they are calculated in case you need detailed background information and explanations
on supported parameters and domains. If you need to understand more of the calculations, if you want to understand what is behind all this e.g. why does the Multiresolution
Segmentation lead to great results or how is distance calculated in eCognition then the
Reference Book is the right choice to work through.
The User Guide is a good starting point to get familiar with workows using eCognition
software. You get to know the world of object based image analysis where classication
and feature extraction offer new possibilities and lead an image analyst to new horizons.
The User Guide is meant as a manual to start using eCognition software realizing simple workows but also advanced classication concepts. Work through the chapters to
improve your image analysis knowledge and get an idea of a new language in image
analysis not only object but also in combination with pixel based image analysis.
For additional help please also refer to our:
User community including e.g. guided tour examples based on data provided, the
latest webinars and the possibility for rule set exchange
(http://www.ecognition.com/community)
Training team offering open trainings, in-company and customized trainings
(http://www.ecognition.com/learn/trainings)
Consultancy unit helping you to solve your image analysis challenges e.g. in feasibility studies or delivering the complete solution development for your projects
(http://www.ecognition.com/products/consulting-services)
Support team (http://www.ecognition.com/support)
Additional geospatial products and services provided by Trimble can be found here:
http://www.trimble.com/imaging/
Installation and licensing guides can be found in the installation directory in the
document InstallationGuide.pdf
This chapter will now introduce some symbols, expressions and basics needed to understand the calculations described in the following chapters of this book.

Introduction

1.1

Symbols and Expressions

1.1.1

Basic Mathematical Notations

Basic mathematical symbols used in expressions.

Therefore

Empty set

a is an element of a set A

b
/B

b is not an element of set B

A B Set A is a proper subset of set B

A 6 B Set A is not a proper subset of set B

A B Set A is a subset of set B

Union of sets A and B

Intersection of sets A and B

A\B

A symmetric difference of sets A and B

The size of set A

Exists, at least one

It follows

For all

Equivalent

ui=1

Sum over index i

[a, b]

Interval with {x | a x b}

1.1.2

Image Layer and Scene

Expressions used to represent image layers and scenes.

k = 1, . . . , K

Image layer k

t = 1, . . . , T

Thematic layer t

(x, y, z,t)

Coordinates of a pixel/voxel

Size of a pixel/voxel in coordinate system unit

(sx, sy, sz, st) Scene extent of a scene s

ck (x, y, z,t)

Intensity value of image layer k at pixel/voxel (x, y, z,t)

cmax
k

Brightest possible intensity value of image layer k

cmin
k

Darkest possible intensity value of image layer k

range
ck

Data range of image layer k

Mean intensity of image layer k

Standard deviation of intensity values of image layer k

N4 (x, y)

4-pixel/voxel neighbors (x, y)

N8 (x, y)

8-pixel/voxel neighbors (x, y)

N6 (x, y, z)

6-pixel/voxel neighbors (x, y, z)

N26 (x, y, z)

26-pixel/voxel neighbors (x, y, z)

Reference Book

11 June 2014

Introduction

1.1.3

Region

Expressions used to represent regions.

Region R

(xG , yG , zG ,tG ) Origin of region R

[Rx , Ry , Rz , Rt ]

Extent of region R

cmax
k (R)
cmin
k (R)

Darkest possible intensity value of image layer k within region R

Brightest possible intensity value of image layer k within region R

ck (R)

Mean intensity of image layer k within region R

k (R)

Standard deviation of image layer k within region R

1.1.4

Pixel Set

Expressions representing layer intensity.

Set of pixels

ck (S)

Mean intensity of image layer k of a set S

k (S)

Standard deviation of intensity values of image layer k of a set S

c(S)

Brightness

wBk
Brightness weight of image layer k
k (v, O) Mean difference of an image object v to image objects in a set O

1.1.5

Image Object

Expressions representing an image object as a set of pixels.

Reference Book

11 June 2014

Introduction

u, v

Image object

Set of image objects

Set of pixels of an image object v

#Pv

Total number of pixels contained in Pv

ck (v)

Mean intensity of image layer k forming an image object v

k (v)

Standard deviation of intensity values of image layer k of all pixels forming image
object v

PvInner

Set of inner border pixels of Pv

PvOuter

Set of outer border pixels of Pv

vi (P)

Image object v in image object level i in pixel P

m(v)

Classication of image object v

Bounding box of an image object v

Bv (d)

Extended bounding box of an image object v with distance d

xmin (v)

Minimum x coordinate of v

xmax (v) Maximum x coordinate of v

ymin (v)

Minimum y coordinate of v

ymax (v) Maximum y coordinate of v

Image object border length

b(v, u)

Length of common border between v and u

1.1.6

Image Objects Hierarchy

Expressions used to represent relations between image objects.

u, v

Image objects

Uv (d)

Superobject of an image object v at a distance d

Sv (d)

Sub-objects of an image object v at a distance d

Vi , i = 1, . . . , n Image object level

Direct neighbors of an image object v

Nv (d)

Neighbors of an image object v at a distance d

e(u, v)

Neighborhood relation between the image objects u and v

1.1.7

Class-Related Set

Expressions used to represent relations between classes.

Reference Book

11 June 2014

Introduction

Set of classes M = {m1, . . . , ma }

A class, (m M)

Nv (d, m) Neighbors classied as m within a distance d

Sv (d, m)

Sub-objects classied as m with hierarchical distance d

Uv (d, m) Superobject classied as m with hierarchical distance d

Vi (m)

All image objects at level i classied as m

(v, m)

Fuzzy membership value of image object v to class m

(v, m)

Stored membership value of image object v to class m

Pi (R, m)

Pixels classied as m on image object level i within region R

1.2

Coordinate Systems Used in eCognition Software

eCognition software uses three coordinate systems:

The pixel coordinate system is used for identifying pixel positions within a map
The user coordinate system allows the use of geocoding information within a map
The internal pixel coordinate system is used only for internal calculations by the
Analysis Engine software.

1.2.1

Pixel Coordinate System

The pixel coordinate system is used to identify pixel position within an image. It is used
for calculating position features such as x -center and y -center.
This coordinate system is oriented from bottom to top and from left to right. The origin
position is (0, 0), which is at the bottom-left corner of the image. The coordinate is
dened by the offset of the bottom-left corner of the pixel from the origin.

Figure 1.1. The pixel coordinate system

Pixel Coordinate Definition

xmin + 1 = xmax ymin + 1 = ymax

Reference Book

11 June 2014

Introduction

Figure 1.2. Coordinates of a pixel

1.2.2

User Coordinate System

The user coordinate system enables the use of geocoding information within a scene. The
values of the separate user coordinate system are calculated from the pixel coordinate
system. In the user interface, the user coordinate system is referred to as the coordinate
system.
This coordinate system is dened by geocoding information:
The bottom-left X position
The bottom-left Y position
Resolution the size of a pixel in coordinate system unit. For example, if the coordinate system is metric, the resolution is the size of a pixel in meters. If the
coordinate system is lat/long, then the resolution is the size of a pixel in degrees
Coordinate system name
Coordinate system type.
The origin of the coordinate system is at the bottom-left corner of the image (x0 , y0 ) .
The coordinate denes the position of the bottom-left corner of the pixel within the user
coordinate system.

Figure 1.3. The user coordinate system

To convert a value from the pixel coordinate system to the user coordinate system and

Reference Book

11 June 2014

Introduction

back, the following transformations are valid, where (x, y) are the coordinates in user
coordinate system and u is the pixel size in units:

1.3

x = x0 + xpixel u
xpixel = (x x0 )/u
y = y0 + ypixel u
ypixel = (y y0 )/u

Image Layer Related Features

1.3.1

Scene

A scene is a collection of combined input image data, as represented in eCognition software. A scene comprises at least one image layer. In addition, it can include more image
layers and thematic layers. You are likely to encounter this concept when importing image or other data into eCognition software.
A scene can also include metadata, such as image creation information or geoinformation, and hold settings such as layer aliases, unit information, or geocoding.
Within eCognition software, image layers, thematic layers, and metadata are loaded by
reference to the respective data les. Each scene is represented by a related map.
A 2D image in a 3D data set is called a slice.
A 2D image in a time series data set is called a frame.
A 4D data set consists of a series of frames where each frame is a 3D data set.
Depending on the related data set, a scene can be one of the following:
Data Set

Scene

2D image

A rectangular area in a 2D space

3D data set

A rectangular volume in a 3D space

4D data set

A series of rectangular volumes in a 4D space

Time series data set A series of rectangular areas in a 2D + time space

Extents

A scene has an origin (x0 , y0 , z0 ,t0 ) , the extent sx in the x -direction, the extent sy in the
y -direction, the extent sz in the z -direction, and the extent st in the t -direction.
The expression is (sx, sy, sz, st) , where:

#(pixels)x is the number of pixels in the x -direction

#(pixels)y is the number of pixels in the y -direction
u is the size of a slice pixel in the co-ordinate system unit
uslices is the spatial distance between slices in the co-ordinate system unit
uframes is the temporal distance between slices in the time co-ordinate unit

Reference Book

11 June 2014

Introduction

The formulas for scene extent are:

sx = #(pixels)x u
sy = #(pixels)x u
sz = #(slices) uslices
st = #(frames) uframes

Coordinates

xpxl is the x offset in pixels

ypxl is the y offset in pixels
zslices is the z offset in slices
t frames is the t offset in frames

Coordinate formulas are:

geo

x = x0 + xpxl u
geo
y = y0 + ypxl u
geo
z = z0 + zslices uslices
geo
t = t0 + t frames uframes

Figure 1.4. Coordinates of a multidimensional scene with four slices and three frames

Layers

Scenes can consist of an arbitrary number of image layers (k = 1, ..., K) and thematic
layers (t = 1, ..., T ) .

1.3.2

Maps

A map represents the combination of a scene and an image object hierarchy. It is the
structure that represents the data that the rule set operates on. eCognition software can
deal with multiple maps and provides algorithms to rescale and copy maps, as well as to
synchronize image objects between them.

Reference Book

11 June 2014

Introduction

1.3.3

Region

A region is a subset denition applicable to maps. A region is specied by its origin,

which is the offset position of the region and its extent [Rx , Ry , Rz , Rt ] in all co-ordinate
dimensions: (xG , yG , zG ,tG ), [Rx , Ry , Rz , Rt ]

Figure 1.5. A three-dimensional region

Depending on map dimensions, the following region types exist:

Type

Expression

(xG , yG ), [Rx , Ry ]

(xG , yG , zG ), [Rx , Ry , Rz ]

(xG , yG , zG ,tG ), [Rx , Ry , Rz , Rt ]

Time series (xG , yG ,tG ), [Rx , Ry , Rt ]

1.3.4

Image Layer

A scene refers to at least one image layer of an image le. Image layers are referenced
by a layer name (string), which is unique within a map.
The pixel/voxel value that is the layer intensity of an image layer k at pixel/voxel (x, y, z,t)
is denoted as ck (x, y, z,t) . The dynamic range of image layers is represented as follows:
cmin
is the smallest possible intensity value of an image layer k
k
cmax
is the largest possible intensity value of an image layer k
k
range
range
ck
is the data range of image layer k with ck
= cmax
cmin
k
k
The dynamic range depends on the image layer data type. The supported image layer
data types 1 are:
1. Full support for image layer data types is dependent on drivers

Reference Book

11 June 2014

Introduction

10
range

Type

cmin
k

cmax
k

8-bit unsigned (int)

255

256

16-bit unsigned (int) 0

65535

65536

16-bit signed (int)

32767

65535

32767

32-bit unsigned (int) 0

32-bit signed (int)
32-bit oat

4294967295 4294967296

2147483647
1.17

10 38

2147483647 4294967295
3.40 1038

n/a

The mean value of all pixel/voxels in a layer is computed by:

ck =

1
ck (x, y)
sx sy (x,y)

The standard deviation of all pixel/voxels in a layer is computed by:

v

u
2
u 1
1
k = t
c
(x,
y)
c
(x,
y)
c
(x,
y)

k
k
k
sx sy (x,y)
sx sy (x,y)
(x,y)
Neighborhood

On raster pixel/voxels there are two ways to dene the neighborhood: as 6-neighborhood
or 26-neighborhood.

1.3.5

Image Layer Intensity on Pixel Sets

A fundamental measurement on a pixel set S and an image object v is the distribution of

the layer intensity. Firstly, the mean intensity within the set is dened by:
ck (S) =

1
ck (x, y)
#s (x,y)S

The standard deviation is dened as:

v

u
u1
1
2
t
k (S) =
(ck (x, y)) #S ck (x, y) ck (x, y)
#S (x,y)S
(x,y)S
(x,y)S

An overall intensity measurement is given by the brightness which is the mean value of
ck (S) for selected image layers.
c(S)
=

1 K B
wk ck (S)
wB k=1

If v is an image object and O a set of other image objects, then the mean difference of the
objects within O to an image object v is calculated by:

1
k (v, 0) = wu ck (v) ck (u)
w u0

Reference Book

11 June 2014

Introduction

1.4
1.4.1

Image Object Related Features

Image Object

An image object is a set of pixels. The set of pixels of an image object v is denoted as Pv .
Image Object Size

At a basic level, the size of an image object is measured by the number of pixels in image
object v. The number of pixels is denoted by #Pv , the cardinality of the respective pixel
set.
Bounding Box of an Image Object

The bounding box Bv of an image object v is the most common form of measurement
of the extent of an image in a scene. It is the smallest rectangular area that encloses all
pixels of v along x- and y-axes. The bounding box is therefore dened by the minimum
and maximum values of the x and y co-ordinates of an image object v xmin (v) , xmax (v)
and ymin (v) , ymax (v) .
The extended bounding box Bv (d) is created by enlarging the bounding box with the same
number of pixels in all directions.

Figure 1.6. Bounding box of an image object v

Pixel Neighborhood and Connectivity of Image Objects

Two pixels are considered to be spatially connected if they neighbor each other in the
pixel raster. On 2D raster images, the neighborhood between pixels can be dened as a
4- or 8-neighborhood:
With reference to the green pixel, the light-blue pixels are the 4-neighborhood. In combination with the dark blue pixels, they form the 8- or diagonal neighborhood of the
green pixel. eCognition software uses the 4-neighborhood for all spatial neighborhood
calculations. As it assumes Image objects to be a set of spatially connected pixels, an

Reference Book

11 June 2014

Introduction

Figure 1.7. 4- and 8-neighborhoods

image object is dened as connected if it is connected according to the 4-neighborhood

principle.
Neighborhood Relations Between Image Objects

Two image objects are considered neighbors if they contain pixels that neighbor each
other according to the 4-neighborhood principle. In other words, two image objects u
and v are considered neighboring each other if this is at least on pixel (x, y) Pv and
one pixel (x , y ) Pu so that (x , y ) is part of N4 (x, y) . The set of all image objects
neighboring v is denoted by Nv (d) :
Nv = {u Vi : (x, y) Pv (x , y ) Pu : (x , y )N4 (x, y)}

Figure 1.8. Neighborhood relation

The entire border line between u and v is called the neighborhood relation and is represented as {$e(u,v).$ The neighborhood relations between image objects are automatically
determined and maintained by the eCognition software.
Border of an Image Object

The border length bv of a two-dimensional image object v is calculated by the number of

the elementary pixel borders. The border length of a three-dimensional image object is
calculated from the borders of image object slices multiplied by the slice distance. The
border of image object slices is counted by the number of the elementary pixel borders.
Similarly, the border length b(v, u) of the neighborhood relation between two image objects v and u is calculated from the common borders of image object slices multiplied
Reference Book

11 June 2014

Introduction

by the slice distance. The border of image object slices is counted by the number of the
elementary pixel borders along the common border. Image objects are basically pixel
sets. The number of pixels belonging to an image object v and its pixel set Pv is denoted
by #Pv .

Figure 1.9. Border length of an image object v or between two objects v , u

The set of all pixels in Pv belonging to the inner border pixels of an image
/ Pv }
object v is dened by PvInner with PvInner = {(x, y) Pv : (x , y ) N4 (x, y) : (x , y )

Inner Border

Figure 1.10. Inner and outer borders of an image object

The set of all pixels in Pv belonging to the outer border pixels of an image
/ Pv : (x , y ) N4 (x, y) : (x , y ) Pv }
object v is dened by PvOuter with PvOuter = {(x, y)
Outer Border

Neighborhood Relations of 3D Image Objects RfB

Image objects in eCognition Developer 9.0 can be two-dimensional or three-dimensional.

The two-dimensional neighborhood concept can be easily extended to three dimensions.
That is, for 3D image objects, there is a 6-neighborhood resulting from the two additional
neighboring pixels in the z dimension. This diagonal 3D neighborhood is also referred to
as the 26-neighborhood.
3D objects are usually 6-connected and may extend across several slices. Several separated parts within a single slice are also possible. However, these parts must be connected
along parts of other slices. If you look at a 3D object in the map view using one of the
three 2D projections, a two-dimensional slice is displayed on the plane. You can use the
MPR view or the 3D view to get a better impression of the shape of a 3D image object.
Reference Book

11 June 2014

Introduction

The software manages two- and three-dimensional neighborhood relations separately.

This allows a correct determination of neighborhood relations in mixed mode situations
where 2D and 3D image objects co-exist in the same scene. 2D image objects can only
be connected by two-dimensional neighborhood relations while 3D image objects can be
connected by both two-dimensional and three-dimensional neighborhood relations.

Figure 1.11. 2D and 3D neighborhood relations

Tracking Image Objects Through Time

There is no special image object type to represent the movement of an image object in
time. Instead, image object links are used to track identical image objects at different
points in time (a time series).
Disconnected Image Objects

In addition to normal 2D and 3D image objects, eCognition Developer 9.0 allows you
to work with image objects that are spatially disconnected. While a connected image
object covers one contiguous region of a scene, a disconnected image object can consist
of arbitrary pixels in two or more potentially disconnected parts within a scene.
It is important to know that image objects can be dened as disconnected even though
they are not disconnected in reality. An image object is dened as disconnected if it lacks
information about its spatial connectivity. The major motivation for this lack of information is the high calculation effort that is necessary to ensure spatial connectivity. If, for
example, you remove some pixels from an image object, it may divide into several parts.
To connect the resulting image objects, a special algorithm needs to analyze the remaining objects and separate them properly into several sub-image objects. If the resulting
image objects remain disconnected, they can be simply marked as there is no need for an

Reference Book

11 June 2014

Introduction

analysis. Disconnected image objects are therefore useful for fast object processing when
spatial information is not relevant. If you use threshold segmentation, for example, to separate an image into two disconnected image objects, you can save a lot of computing time
and memory since the number of image objects is considerably reduced.

Figure 1.12. Disconnected image objects (left) and spatially connected image object (right)

The main purpose of working with disconnected image objects is the representation of
simple pixel crowds, where several properties that can be measured for normal image
objects might not make sense or are meaningless. You should therefore note that the
following information is not available for disconnected image objects:
Neighborhood relations and all features based on these relations
Shape features
Polygons.

1.4.2

Image Object Hierarchy

Image objects are organized into levels, where each object on each level creates a partition
of the scene S . This can be described using the following fundamental conditions. The
totality of all image objects covers the entire scene:
U Pv = {(x, y)}

vVi

There are no image objects that overlap (having a non-empty intersection):

Pu Pv =

u,vVi

The image object levels are hierarchically structured. This means that all image objects
on a lower level are complete contained in exactly one image object of a higher level.

i< j

vVi

uVi

Reference Book

11 June 2014

Introduction

Level Distance

The level distance represents the hierarchical distance between image objects on different
levels in the image object hierarchy. Starting from the current image object level, the
level distance indicates the hierarchical distance of image object levels containing the
respective image objects (sub-objects or superobjects).
Since each object has exactly 1 or 0 superobjects on the higher level, the superobject of
v with a level distance d can be denoted as Uv (d). Similarly, all sub-objects with a level
distance d are denoted as Sv (d).

Figure 1.13. Image object hierarchy

Spatial Distance

The spatial distance represents the distance between image objects on the same level in
the image object hierarchy. If you want to analyze neighborhood relations between image
objects on the same image object level in the image object hierarchy, the feature distance
expresses the spatial distance (in pixels) between the image objects. The default value is
0 (in other words, only neighbors that have a mutual border are considered). The set of
all neighbors within a distance d are denoted by Nv (d).
Many features enable you to enter a spatial distance parameter. Distances are usually measured in pixel units. Because exact distance measurements
between image objects are very processor-intensive, eCognition software uses approximation approaches to estimate the distance between image objects.
Distance Measurements

There are two different approaches: the center of gravity and the smallest enclosing rectangle. You can congure the default distance calculations.
The Center of Gravity approximation measures the distance between
the center of gravity between two image objects. This measure can be computed very
efciently but it can be inaccurate for large image objects.
Center of Gravity

Reference Book

11 June 2014

Introduction

The smallest enclosing rectangle approximation tries to

correct the center of gravity approximation by using rectangular approximations of the
image object to adjust the basic measurement delivered by the center of gravity.
Smallest Enclosing Rectangle

Figure 1.14. Distance calculation between image objects. (The black line is the center of

gravity approximation. The red line is the smallest enclosing rectangle approximation)

We recommend using the center of gravity distance for most applications although the
smallest enclosing rectangle may give more accurate results. A good strategy for exact
distance measurements is to use center of gravity and try to avoid large image objects, for
example, by creating border objects. To avoid performance problems, restrict the total
number of objects involved in distance calculations to a small number.
You can edit the distance calculation in the algorithm parameters of the Set Rule Set
Options algorithm and set the Distance Calculation option to your preferred value.

1.5
1.5.1

Class-Related Features
Class-Related Sets

Let M = {m1, . . . , ma } be a set of classes with m being a specic class and m M. Each
image object has a fuzzy membership value of (v, m) to class m . In addition each
image object also carries the stored membership value (u, m) that is computed during
execution of the last classication algorithm. By restricting a set of image objects O
to only the image object that belong to class m , various class-related features can be
computed:
m) = 1} Neighbors of class
Nv (d, m) = {u Nv (d) : (u,

m) = 1}
Sv (d, m) = {u Sv (d) : (u,

Sub-objects of class

m) = 1}
Vi (m) = {u Vi (m) : (u,

Image object levels of class

m) = 1} Superobjects of class
Uv (d, m) = {u Uv (d) : (u,
Pi (R, m) = {p R : (vi (p)) = m}

Pixels of class

Example: The mean difference of layer k to a neighbor image object within a distance d
and that image object belongs to a class m is dened as k (v, Nv (d, m)).

Reference Book

11 June 2014

Introduction

1.6

Shape-Related Features

Many of the eCognition form features are based on the statistics of the spatial distribution
of the voxels that form an 3D image object. As a central tool to work with these statistics,
eCognition Developer 9.0 uses the covariance matrix:

1.6.1

Parameters

X : x co-ordinates of all voxels forming the image object

Y : y co-ordinates of all voxels forming the image object
Z : z co-ordinates of all voxels forming the image object.

1.6.2

Expression

Var(X) Cov(XY ) Cov(XZ)

C = Cov(XY ) Var(Y )
Cov(Y Z)
Cov(XZ) Cov(Y Z)
Var(Z)

Another frequently used technique to derive information about the form of image objects
is the bounding box approximation. Such a bounding box can be calculated for each
image object and its geometry can be used as the rst clue to the image object itself.
The main information provided by the bounding box is its length a , its width b , its
thickness c , its volume a b c and its degree of lling f , which is the volume V lled
by the image object divided by the total volume a b c of the bounding box.

1.6.3

Shape Approximations Based on Eigenvalues

The shape approximations based on eigenvalues measures the statistical distribution of

the voxel co-ordinates (x, y, z) of a set Pv .
The center of gravity of the set Pv is:
xcenter =

1
x
#Pv (x,y,z)

ycenter =

1
y
#Pv (x,y,z)

zcenter =

1
z
#Pv (x,y,z)

The variances of voxel co-ordinates are:

2

1
1
2
Cxx =
xi #P v xi = Exx Ex2
#Pv (x,y,z)
(x,y,z)
Cyy =

2
1
1
2
y

y
i #P v i = Eyy Ey2
#Pv (x,y,z)
(x,y,z)

Reference Book

11 June 2014

Introduction

2
1
1
2
y

y
i #P v i = Ezz Ez2
#Pv (x,y,z)
(x,y,z)

1
1
1
x
y

x
y
Cxy =

i i #Pv i
i = Exy Ex Ey
#Pv (x,y,z)
#Pv (x,y,z)
(x,y,z)

1
1
1
Cxz =
xi zi #Pv xi #Pv zi = Exz Ex Ez
#Pv (x,y,z)
(x,y,z)
(x,y,z)

1
1
1
Cyz =
yi zi #Pv yi #Pv zi = Eyz Ey Ez
#Pv (x,y,z)
(x,y,z)
(x,y,z)
Czz =

The covariance matrix C of voxel co-ordinates is:

Cxx Cxy Cxz

C = Cyx Cyy Cyz

Czx Czy Czz

The diagonalization of the covariance matrix gives three eigenvalues (1 , 2 , 3 ) , which

are the main, medium, and minor axes of an ellipsoid.
e = C e , where is the eigenvalue and e is the eigenvector.

1.6.4

Elliptic Approximation

The elliptic approximation uses the eigenvalues (1 , 2 , 3 ) of the covariance matrix and
computes an ellipsis with axis along the eigenvector e1 with length a , and along the
eigenvector e2 with length b , and along the eigenvector e3 with length c .
The formula is a : b : c = 1 : 2 : 3 .

Figure 1.15. Elliptic approximation

Reference Book

11 June 2014

Introduction

1.7

About Algorithms

A single process executes an algorithm on a domain. It is the elementary unit of a rule

set and provides a solution to a specic image analysis problem. Processes are the main
working tools for developing rule sets. A rule set is a sequence of processes that are
executed in a dened order.
A domain is for instance a set of image objects. Every process loops through this set of
image objects one by one and applies the algorithm to each single image object. This
image object is referred to as the current image object.

1.7.1

Creating a Process

A single process can be created using the Edit Process dialog box, in which you can
dene:
The method of the process from an algorithm list, for example the Multiresolution
Segmentation algorithm or a Classication algorithm.
The domain on which an algorithm should be performed, for example the pixel
level, an image object level or a vector layer.
Detailed algorithm parameter settings varying dependent on the algorithm selected.

Figure 1.16. The Edit Process dialog box

1.7.2

Specifying Algorithm Parameters

Depending on the chosen algorithm, several parameters may be modied:

1. Dene the individual settings of the algorithm in the Algorithms Parameters group
box. If available, click a plus sign (+) button to expand the table to access additional
parameters.
Reference Book

11 June 2014

Introduction

2. To edit values of parameters, select the parameter name or its value by clicking on
it. Depending on the type of value, change the value in one of the following ways:
Edit values directly within the value eld.
Click the ellipsis button located inside the value eld. A dialog box opens,
enabling you to congure the value.
Select the value from a drop-down list. For many parameters, you can select
a feature by clicking the From feature item. or an existing variable. Alternatively, you can select a variable. To create a new variable, type a name for the
new variable and click OK, or press Enter to open the Create Variable dialog
box for additional settings.

Reference Book

11 June 2014

2 Process-Related Operation
Algorithms
The Process-Related Operation algorithms are used to control other processes.

2.1

Execute Child Processes

Execute all child (subordinate) processes of a process. To dene a process using this
algorithm as a parent process, you can right-click it and choose the Insert Child command
in the context menu.

2.1.1

Supported Domains

Execute

Use the Execute Child Processes algorithm in conjunction with the Execute domain to
structure your process tree. A process with this setting serves as a parent process, providing a container for a sequence of functionally related child processes.
Pixel Level

Applies the algorithm at the pixel level. Typically used for initial segmentations.
Image Object Level

Use the Execute Child Processes algorithm in conjunction with other domains (for example, the image object level domain) to loop over a set of objects. All contained child
processes will be applied to the objects in the respective domain. In this case the child
processes usually use one of the following as image object domain: current image object,
neighbor object, super object, sub objects.
Current Image Object

Applies the algorithm to the current internally selected image object of the parent process.

Process-Related Operation Algorithms

Neighbor Image Object

Applies the algorithm to all neighbors of the current internally selected image object of
the parent process. The size of the neighborhood is dened by the Distance parameter.
Super Object

Applies the algorithm to the superobject of the current internally selected image object
of the parent process. The number of levels above in the image object level hierarchy is
dened by the Level Distance parameter.
Sub Object

Applies the algorithm to the sub objects of the current internally selected image object
of the parent process. The number of levels below in the image object level hierarchy is
dened by the Level Distance parameter.
Maps

Applies the algorithm to all specied maps of a project. You can select this domain in
parent processes with the Execute Child Process algorithm to set the context for child
processes that use the map parameter From Parent.
Linked Objects

Applies the algorithm to all linked image objects of the current internally selected image
object of the parent process.
Image Object List

Applies the algorithm to all image objects that were collected with the algorithm Update
Image Object List.
Array

Options for the array function are visible when Array is selected as Domain. For more
information on arrays, please consult also the eCognition Developer User Guide > Advanced Rule Set Concepts > Arrays and have a look at the examples in our user community:
http://community.ecognition.com/home/Arrays%20Example%20%231.zip/view or http:
//community.ecognition.com/home/Arrays%20Example%20%232.zip/view
Array: Select a predened array
Array Type: Displays the array type (this is not editable)
Variable: The variable holding the value of the current array item when the process
is executed

Reference Book

11 June 2014

Process-Related Operation Algorithms

Index Variable: The scene variable holding the index of the current array item
when the process is executed
Vectors

Applies the algorithm to a thematic vector layer that can be selected in Domain > Parameter > Thematic Vector Layer.
Vectors (Multiple layers)

Applies the algorithm to a set of thematic vector layers that can be selected in the Select
thematic layers dialog window opened by clicking the ellipsis button located inside the
value eld of Domain > Parameter > Thematic Vector Layer.

2.2

Execute Child As Series

When items are exported during a multimap analysis they are held in memory, which can
often lead to out-of-memory problems during intensive analyses. The Execute Child as
Series algorithm allows partial results to be exported as soon as they are available, freeing
up memory for other tasks.

2.2.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Sub Object; Super Object; Linked Objects; Maps; Image Object List

2.2.2

Algorithm Parameters

Series Name

Enter the series name variables and strings are possible. All export algorithms receive a
new parameter: Export Series (Yes/No)

2.3

If, Then and Else

If the conditions in the If algorithm are connected, all the dened conditions must be true
to enter the Then path, otherwise the Else path is taken
If no condition is dened, the Then path is chosen
If there is more than one Then or Else condition, all are executed.

Reference Book

11 June 2014

Process-Related Operation Algorithms

2.3.1

Supported Domains

Execute; Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub Objects; Linked Objects; Maps; Image Object List; Array

2.4

Throw

The Throw algorithm generates an error message (the Exception Message parameter) if a
certain condition occurs, for example the number of objects exceeds a dened number.

2.4.1

Supported Domains

Execute; Pixel Level; Image Object Level: Current Image Object; Neighbor Image object;
Super Object; Sub Objects; Linked Objects; Maps; Image Object List; Array

2.5

Catch

If an exception (generating an error message) occurs as dened in the Throw algorithm

and a Catch algorithm has been dened, the error message is not generated. Instead the
processes (and any child processes) dened in the Catch algorithm are executed. If there
are Catch processes without an exception, the Catch process is not executed.

2.5.1

Supported Domains

Execute; Pixel Level; Image Object Level: Current Image Object; Neighbor Image object;
Super Object; Sub Objects; Linked Objects; Maps; Image Object List; Array

2.6

Set Rule Set Options

The Set Rule Set Options algorithm lets you control certain settings for the rule set, or
parts of rule sets. For example, you may want to apply particular settings to analyze large
objects and change them to analyze small objects. In addition, because the settings are
part of the rule set and not on the client, they are preserved when the rule set is run on a
server.
You can save the rule set or the associated project to preserve the settings chosen in this
algorithm, because saving the project also saves the rule set. The current settings can be
displayed under Tools - Options - Project Settings.

2.6.1

Supported Domains

Execute

Reference Book

11 June 2014

Process-Related Operation Algorithms

2.6.2

Algorithm Parameters

Apply to Child Processes Only

If the value is No, settings apply globally, persisting after completion of execution. If Yes
is selected, changes apply only to child processes.
Distance Calculation

Smallest Enclosing Rectangle uses the smallest enclosing rectangle of an image

object for distance calculations
Center of Gravity uses the center of gravity of an image object for distance calculations
Default reverts values to their defaults when the rule set is saved
Keep Current maintains the existing settings when the rule set is saved
Current Resampling Method

Center of Pixel initiates resampling from the center of the pixel.

Upper Left Corner of Pixel initiates resampling from the upper-left corner of the
pixel.
Default reverts values to their defaults when the rule set is saved
Keep Current maintains the existing settings when the rule set is saved
Evaluate Conditions on Undefined Features as 0

When a value is undened, with respect to a condition you have specied, the software
can evaluate it as false, or perform the evaluation based on a value of zero.
Yes: Evaluate any condition with undened feature as false (this is the default).
No: Assign value zero to undened feature before evaluation of any condition
with undened feature.
Default reverts values to their defaults when the rule set is saved
Keep Current maintains the existing settings when the rule set is saved.
Polygons Base Polygon Threshold

This value sets the degree of abstraction for the base polygons. The default is 1.25.
Polygons Shape Polygon Threshold

This value determines the degree of abstraction for the shape polygons. Shape polygons
are independent of the topological structure and consist of at least three points. The
threshold for shape polygons can be changed any time without the need to recalculate the
base vectorization. The default value is one.

Reference Book

11 June 2014

Process-Related Operation Algorithms

Polygons Remove Slivers

Remove Slivers is used to avoid intersection of edges of adjacent polygons and selfintersections of polygons.
Sliver removal becomes necessary with higher threshold values for base polygon generation. Note that the processing time to remove slivers is high, especially for low thresholds
where it is not needed anyway.
No: enables the intersection of polygon edges and self-intersections
Yes: avoids the intersection of edges of adjacent polygons and self-intersections of
polygons
Default reverts values to their defaults when the rule set is saved
Keep Current maintains the existing settings when the rule set is saved
Update Topology

Update Topology allows you to update neighborhood relations automatically or on demand.

Save Temporary Layers

Selecting Yes saves temporary layers to your hard drive. The default values in eCognition
Developer 9.0 are:
Yes: for new rule sets.
No: for rule sets created with eCognition 8.8 or earlier.
Yes: for rule sets created with eCognition 9.0.
Polygon Compatibility Mode

Set this option to use polygon compatibility option for polygons.

Resampling Compatibility Mode

Set this option to use the compatibility option for image layer resampling. After changing
this value the image objects may need to be recreated to reect the changes.
Switch to classification view after process execution

This parameter switches to the classication view after process execution.

Point cloud distance filter

This parameter sets the maximum laser point distance to the sensor. The distance unit is
according to the unit of the point cloud projection.

Reference Book

11 June 2014

3 Segmentation Algorithms
Segmentation algorithms are used to subdivide entire images at a pixel level, or specic
image objects from other domains into smaller image objects.
Trimble provides several different approaches to segmentation, ranging from very simple
algorithms, such as chessboard and quadtree-based segmentation, to highly sophisticated
methods such as multiresolution segmentation and contrast lter segmentation.
Segmentation algorithms are required whenever you want to create new image objects
levels based on image layer information. But they are also a very valuable tool to rene
existing image objects by subdividing them into smaller pieces for more detailed analysis.

3.1

Chessboard Segmentation

The Chessboard Segmentation algorithm splits the pixel domain or an image object domain into square image objects.
A square grid aligned to the image left and top borders of xed size is applied to all
objects in the domain and each object is cut along these gridlines.

Figure 3.1. Result of chessboard segmentation with object size 20

Segmentation Algorithms

3.1.1

Supported Domains

Pixel Level; Image Object Level: Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

3.1.2

Algorithm Parameters

Object Size

Object Size denes the size of the square grid in pixels. Variables are rounded to the
nearest integer
Level Name

In the Level Name eld, enter the name of a new image object level. This parameter is
only available, if the domain Pixel Level is selected in the process dialog.
Overwrite Existing Level

This parameter is only available when Pixel Level is selected. It allows you to automatically delete an existing image level above the pixel level and replace it with a new level
created by the segmentation.
Thematic Layer Usage

In the Thematic Layers eld, specify the thematic layers to be considered in addition
to segmentation. Each thematic layer used for segmentation will cause further splitting
of image objects while enabling consistent access to its thematic information. You can
segment an image using more than one thematic layer. The results are image objects representing proper intersections between the thematic layers. If you want to produce image
objects based exclusively on thematic layer information, you can select a chessboard size
larger than your image size.

3.2

Quadtree-Based Segmentation

The Quadtree-Based Segmentation algorithm splits the pixel domain or an image object
domain into a quadtree grid formed by square objects.
A quadtree grid consists of squares with sides each having a power of two and aligned
to the image left and top borders. It is applied to all objects in the domain and each
object is cut along these gridlines. The quadtree structure is built so that each square has
a maximum possible size and also fullls the homogeneity criteria dened by the mode
and scale parameters.
The maximum square object size is 256 x 256, or 65,536 pixels.

Reference Book

11 June 2014

Segmentation Algorithms

Figure 3.2. Result of quadtree-based segmentation with mode color and scale 40

3.2.1

Supported Domains

Pixel Level; Image Object Level: Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

3.2.2

Algorithm Parameters

Mode

Color: The maximal color difference within each square image object is less than
the Scale value
Super Object Form: Each square image object must completely t into the superobject. This mode only works with an additional upper image level
Scale

Scale denes the maximum color difference within each selected image layer inside
square image objects. It is only used in conjunction with the Color mode.
Level Name

In the Level Name eld, enter the name of a new image object level. This parameter is
only available if the domain Pixel Level is selected in the process dialog.
Overwrite Existing Level

Reference Book

11 June 2014

Segmentation Algorithms

Image Layer Weights

Image layers can be weighted depending on their importance or suitability for the segmentation result. The higher the weight assigned to an image layer, the more weight
will be assigned to that layers pixel information during the segmentation process, if it is
used. Consequently, image layers that do not contain the information intended for representation by the image objects should be given little or no weight. For example, when
segmenting a geographical LANDSAT scene using multiresolution segmentation or spectral difference segmentation, the segmentation weight for the spatially coarser thermal
layer should be set to 0 in order to avoid deterioration of the segmentation result by the
blurred transient between image objects of this layer.
1. In the Algorithm parameters area, expand the Image Layer weights list and set the
weight of the image layers to be considered by the algorithm. You can use both of
the following methods:
Select an image layer and edit the weight value
Select Image Layer weights and click the ellipsis button located inside the
value eld to open the Image Layer Weights dialog box
2. Select an image layer in the list. To select multiple image layers press Ctrl.
3. Enter a new weight in the New value text box and click Apply.

Figure 3.3. Image Layer Weights dialog box

Options

Click the Calculate Stddev button to check the image layer dynamics. The calculated standard deviations of the image layer values for each single image layer are
listed in the Stddev. column.
To search for a specic layer, type the name into the Find text box.

Reference Book

11 June 2014

Segmentation Algorithms

3.2.3

Thematic Layer Weights

In the Thematic Layers eld, specify the thematic layers to be considered in addition to
segmentation. Each thematic layer used for segmentation will cause additional splitting
of image objects while enabling consistent access to its thematic information. You can
segment an image using more than one thematic layer. The results are image objects representing proper intersections between the thematic layers. If you want to produce image
objects based exclusively on thematic layer information, you can select a chessboard size
larger than your image size between the thematic layers.

3.3

Contrast Split Segmentation

The Contrast Split Segmentation algorithm segments an image or image object into dark
and bright regions. It is based on a threshold that maximizes the contrast between the
resulting bright objects (consisting of pixels with pixel values above the threshold) and
dark objects (consisting of pixels with pixel values below the threshold).
The algorithm evaluates the optimal threshold separately for each image object in the domain. If the pixel level is selected in the domain, the algorithm rst executes a chessboard
segmentation, then performs the split on each square.
It achieves the optimization by considering different pixel values as potential thresholds.
The test thresholds range from the minimum threshold to the maximum threshold, with
intermediate values chosen according to the step size and stepping type parameter. If
a test threshold satises the minimum dark area and minimum bright area criteria, the
contrast between bright and dark objects is evaluated. The test threshold causing the
largest contrast is chosen as the best threshold and used for splitting.

3.3.1

Supported Domains

Pixel Level; Image Object Level: Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

3.3.2

Settings

Chessboard Tile Size

This eld is available only if pixel level is selected in the Domain. Enter the chessboard
tile size (the default is 1,000).
Level Name

Select or enter the level to contain the results of the segmentation. Available only if the
pixel level is selected in the domain.

Reference Book

11 June 2014

Segmentation Algorithms

Overwrite Existing Level

Enter the minimum gray value to be considered for splitting. The algorithm calculates the
threshold for gray values from the minimum threshold value to the maximum threshold
value (the default is 0).
Maximum Threshold

Enter the maximum gray value to be considered for splitting. The algorithm calculates the
threshold for gray values from the minimum threshold value to the maximum threshold
value (the default is 255).
Step Size

Enter the step size by which the threshold increases from the minimum threshold to the
maximum threshold. The value is either be added to the threshold or multiplied by the
threshold, according to the selection in the Stepping Type eld. The algorithm recalculates a new best threshold each time the threshold is changed by application of the values
in the Step Size and Stepping Type elds, until the maximum threshold is reached. Higher
values entered for step size tend to execute more quickly; smaller values tend to achieve
a split with a larger contrast between bright and dark objects.
Stepping Type

Use the drop-down list to select one of the following:

Add: Calculate each step by adding the value in the Scan Step eld
Multiply: Calculate each step by multiplying by the value in the Scan Step eld
Image Layer

Select the image layer where the contrast is to be maximized.

Class for Bright Objects

Create a class for image objects brighter than the threshold or select one from the dropdown list. Image objects are not classied if the value in the Execute splitting eld is
No.

Reference Book

11 June 2014

Segmentation Algorithms

Class for Dark Objects

Create a class for image objects darker than the threshold or select one from the dropdown list. Image objects are not classied if the value in the Execute splitting eld is
No.

3.3.3

Advanced Settings

Contrast Mode

Select the method the algorithm uses to calculate contrast between bright and dark objects.
The algorithm calculates possible borders for image objects (where a is the mean of bright
border pixels and b is the mean of dark border pixels):
Edge Ratio: (a b) (a + b)
Edge Difference: (a b)
Object Difference: The difference between the mean of all bright pixels and the
mean of all dark pixels.
Execute Splitting

Select Yes to split objects with the best-detected threshold. Select No to simply compute
the threshold without splitting.
Variable for Best Threshold

Enter a scene variable to store the computed pixel value threshold that maximizes the
contrast.
Variable for Best Contrast

Enter a scene variable to store the computed contrast between bright and dark objects
when splitting with the best threshold. The computed value is different for each contrast
mode.
Minimum Relative Area Dark

Enter the minimum relative dark area. Segmentation into dark and bright objects only
occurs if the relative dark area is higher than the value entered. Only thresholds that
lead to a relative dark area larger than the value entered are considered as best thresholds.
Setting this value to a number greater than 0 may increase speed of execution.
Minimum Relative Area Bright

Enter the minimum relative bright area. Only thresholds that lead to a relative bright area
larger than the value entered are considered as best thresholds. Setting this value to a
number greater than 0 may increase speed of execution.
Reference Book

11 June 2014

Segmentation Algorithms

Minimum Contrast

Enter the minimum contrast value threshold. Segmentation into dark and bright objects
only occurs if a contrast higher than the value entered can be achieved.
Minimum Object Size

Enter the minimum object size in pixels that can result from segmentation. Only larger
objects are segmented. Smaller objects are merged with neighbors randomly. (The default
value of 1 effectively inactivates this option.)

3.4

Multiresolution Segmentation

The Multiresolution Segmentation algorithm locally minimizes the average heterogeneity

of image objects for a given resolution of image objects. It can be executed on an existing
image object level or the pixel level for creating new image objects on a new image object
level.

Figure 3.4. Result of multiresolution segmentation with scale 10, shape 0.1 and compactness

0.5

The multiresolution segmentation algorithm consecutively merges pixels or existing image objects. Thus it is a bottom-up segmentation algorithm based on a pairwise region
merging technique. Multiresolution segmentation is an optimization procedure which,
for a given number of image objects, minimizes the average heterogeneity and maximizes
their respective homogeneity.
The segmentation procedure works according the following rules, representing a mutualbest-tting approach:
1. The segmentation procedure starts with single image objects of one pixel and repeatedly merges them in several loops in pairs to larger units as long as an upper
threshold of homogeneity is not exceeded locally. This homogeneity criterion is
dened as a combination of spectral homogeneity and shape homogeneity. You

Reference Book

11 June 2014

Segmentation Algorithms

2.
3.
4.
5.
6.

can inuence this calculation by modifying the scale parameter. Higher values for
the scale parameter result in larger image objects, smaller values in smaller image
objects.
As the rst step of the procedure, the seed looks for its best-tting neighbor for a
potential merger.
If best tting is not mutual, the best candidate image object becomes the new seed
image object and nds its best tting partner.
When best tting is mutual, image objects are merged.
In each loop, every image object in the image object level will be handled once.
The loops continue until no further merger is possible.

Figure 3.5. Each image object uses the homogeneity criterion to determine the best neighbor

to merge with

Figure 3.6. If the first image objects best neighbor (red) does not recognize the first image

object (gray) as best neighbor, the algorithm moves on (red arrow) with the second image
object finding the best neighbor

Figure 3.7. This branch-to-branch hopping repeats until mutual best fitting partners are

found

The procedure continues with another image objects best neighbor. The procedure iterates until no further image object mergers can be realized without violating the maximum
allowed homogeneity of an image object.
With any given average size of image objects, multiresolution segmentation yields good
abstraction and shaping in any application area. However, it has higher memory requirements and signicantly slower performance than some other segmentation techniques and
therefore is not always the best choice.
Reference Book

11 June 2014

Segmentation Algorithms

Figure 3.8. If the homogeneity of the new image object does not exceed the scale parameter,

the two partner image objects are merged.

3.4.1

Supported Domains

Pixel level; Image Object Level

3.4.2

Level Settings

Level Name

The Level Name eld lets you dene the name for the new image object level. It is only
available if a new image object level will be created by the algorithm. To create new
image object levels, use either the domain Pixel Level in the process dialog or set the
Level Usage parameter to Create Above or Create Below.
Overwrite Existing Level

Select one of the available modes from the drop-down list. The algorithm is applied
according to the mode based on the image object level that is specied by the domain.
This parameter is not visible if pixel level is selected as domain in the Edit Process dialog
box.
Use Current applies multiresolution segmentation to the existing image object level.
Objects can be merged and split depending on the algorithm settings
Use Current (Merge Only) applies multiresolution segmentation to the existing image object level. Objects can only be merged. Usually this mode is used together
with stepwise increases of the scale parameter
Create Above creates a copy of the image object level as super objects
Create Below creates a copy of the image object level as sub objects.

Reference Book

11 June 2014

Segmentation Algorithms

3.4.3

Segmentation Settings

Image Layer Weights

Image layers can be weighted depending on their importance or suitability for the segmentation result. The higher the weight assigned to an image layer, the more weight
will be assigned to that layers pixel information during the segmentation process, if it is
used. Consequently, image layers that do not contain the information intended for representation by the image objects should be given little or no weight. For example, when
segmenting a geographical LANDSAT scene using multiresolution segmentation or spectral difference segmentation, the segmentation weight for the spatially coarser thermal
layer should be set to 0 in order to avoid deterioration of the segmentation result by the
blurred transient between image objects of this layer.
In the Algorithm parameters area, expand the Image Layer weights list and set the
weight of the image layers to be considered by the algorithm. Select an image layer
and edit the respective weight value.
Alternatively you can insert the weights in the value eld of the parameter Image
layer weights, separated by a comma, e.g.: 0,0,1
You can also use a variable as a layer weight.
Compatibility mode

Compatibility mode to previous software versions (eCognition Version 8.64 or earlier)

where a separate dialog Image Layer Weights could be opened by clicking the ellipsis
button located inside the value eld of the image layer weights parameter (select Domain
- image object level and Parameter Level usage - Create below).
Thematic Layer Usage

Specify the thematic layers to be candidates for segmentation. Each thematic layer that
is used for segmentation will lead to additional splitting of image objects while enabling
consistent access to its thematic information. You can segment an image using more
than one thematic layer. The results are image objects representing proper intersections
between the thematic layers.
Scale Parameter

The Scale Parameter is an abstract term that determines the maximum allowed heterogeneity for the resulting image objects. For heterogeneous data, the resulting objects for
a given scale parameter will be smaller than in more homogeneous data. By modifying
the value in the Scale Parameter value you can vary the size of image objects.

Reference Book

11 June 2014

Segmentation Algorithms

TIP: Always produce image objects of the biggest possible scale that still
distinguish different image regions (as large as possible and as ne as necessary). There is a tolerance concerning the scale of the image objects
representing an area of a consistent classication due to the equalization
achieved by the classication. The separation of different regions is more
important than the scale of image objects.

3.4.4

Composition of Homogeneity Criterion

The object homogeneity to which the scale parameter refers is dened in the Composition
of Homogeneity criterion eld. In this circumstance, homogeneity is used as a synonym
for minimized heterogeneity. Internally, three criteria are computed: color, smoothness,
and compactness. These three criteria for heterogeneity may be applied in many ways
although, in most cases, the color criterion is the most important for creating meaningful
objects. However, a certain degree of shape homogeneity often improves the quality
of object extraction because the compactness of spatial objects is associated with the
concept of image shape. Therefore, the shape criteria are especially helpful in avoiding
highly fractured image object results in strongly textured data (for example radar data).

Figure 3.9. Multiresolution concept flow diagram

Shape

The value of the Shape eld modies the relationship between shape and color criteria; By
modifying the Shape criterion, 1 you dene the color criteria (color = 1 shape). In effect,
by decreasing the value assigned to the Shape eld, you dene to which percentage the
spectral values of the image layers will contribute to the entire homogeneity criterion.
1. A high value for the shape criterion operates at the cost of spectral homogeneity. However, spectral information
is in the end the primary information contained in image data. Using the shape criterion too intensively may
thus reduce the quality of segmentation results.

Reference Book

11 June 2014

Segmentation Algorithms

This is weighted against the percentage of the shape homogeneity, which is dened in the
Shape eld.
Changing the weight for the Shape criterion to 1 will result in objects more optimized for
spatial homogeneity. However, the shape criterion cannot have a value larger than 0.9,
due to the fact that without the spectral information of the image, the resulting objects
would not be related to the spectral information at all. The slider bar adjusts the amount
of Color and Shape to be used for the segmentation.
In addition to spectral information, the object homogeneity is optimized with regard to
the object shape, dened by the Compactness parameter.
Compactness

The compactness criterion is used to optimize image objects with regard to compactness.
This criterion should be used when different image objects which are rather compact, but
are separated from non-compact objects only by a relatively weak spectral contrast. Use
the slider bar to adjust the degree of compactness to be used for the segmentation.

3.5

Spectral Difference Segmentation

The Spectral Difference Segmentation algorithm merges neighboring image objects according to their mean image layer intensity values. Neighboring image objects are
merged if the difference between their layer mean intensities is below the value given
by the maximum spectral difference.
This algorithm is designed to rene existing segmentation results, by merging spectrally
similar image objects produced by previous segmentations. It cannot be used to create
new image object levels based on the pixel level domain.

3.5.1

Supported Domains

Image Object Level

3.5.2

Level Settings

Level Usage

Dene the image object level to be used: the current one or a new one to be created above
the current.
Level Name

Dene the Level Name of the image object level to be created. Not available if you use
the current image object level.

Reference Book

11 June 2014

Segmentation Algorithms

3.5.3

Segmentation Settings

Maximum Spectral Difference

Dene the maximum spectral difference in gray values between image objects that are
used during the segmentation. If the difference is below this value, neighboring objects
are merged.
Parameters

wn[k1,k2,k3] are the normalized layer weights w[k1,k2,k3]

wk1
(wk1 + wk2 + wk3 )
wk2
wnk2 =
(wk1 + wk2 + wk3 )
wk3
wnk3 =
(wk1 + wk2 + wk3 )
wnk1 =

Expression

wnk1 abs(k11 k12 ) + wnk2 abs(k21 k22 ) + wnk3 abs(k31 k32 )

Image Layer Weights

Enter weighting values the higher the weight assigned to an image layer, the more
weight will be given to that layers pixel information during the segmentation process.
You can also use a variable as a layer weight.
Thematic Layer Usage

Specify the thematic layers that are to be considered in addition for segmentation. Each
thematic layer used for segmentation will lead to additional splitting of image objects,
while enabling consistent access to its thematic information. You can segment an image
using more than one thematic layer. The results are image objects representing proper
intersections between the thematic layers.

3.6

Multi-Threshold Segmentation

Multi-Threshold Segmentation splits the domain based on pixel values. This creates image objects and classies them based on user-created thresholds. It can also be used to
create unclassied image objects based on pixel values thresholds.

Reference Book

11 June 2014

Segmentation Algorithms

3.6.1

Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

3.6.2

Level Settings

Level Name

Available only if the domain is pixel level. Select an existing level or enter a name.
Overwrite Existing Level

Select an image layer to use for threshold evaluation.

Ensure Connected Objects

Select the type of image objects that are created (the domain must be pixel level):
If Yes, all image objects are marked as connected
If No, image objects can be marked as disconnected
Merge Image Objects First

This parameter species the behaviour if the Ensure Connected Objects parameter is set
to Yes. It is only available if you select a domain that represents existing image objects.
If Yes, all image objects with the same classication are rst merged into one
disconnected object and then separated again into connected image objects. By
using this option you might lose some borders between image objects that existed
before executing the algorithm.
If No, only disconnected image objects are further separated into connected image
objects where necessary.
Min Object Size

Limits the minimum size of resulting image objects to a dened value.

Reference Book

11 June 2014

Segmentation Algorithms

Thresholds

You can set multiple thresholds to classify image objects based on pixel values. An
additional threshold eld is added as each threshold is created. Multiple thresholds must
be in ascending order:
Class 1

Pixels below the threshold will be classied as the class selected or entered.

Enter a pixel value below which pixels will be classied as dened in the
Class eld. Alternatively, you can select a feature or a variable. To create a new variable,
type a name for the new variable and click OK to open the Create Variable dialog box.
Threshold 1

3.7

Contrast Filter Segmentation

Pixel lters detect potential objects by contrast and gradient, and create suitable object
primitives. An integrated reshaping operation modies the shape of image objects to help
form coherent and compact image objects. The resulting pixel classication is stored in
an internal thematic layer. Each pixel is classied as one of the following classes: No
Object, Object in First Layer, Object in Second Layer, Object in Both Layers and Ignored
by Threshold. Finally a chessboard segmentation is used to convert this thematic layer
into an image object level.
Contrast lter segmentation, as a rst step, improves overall image analysis performance
substantially.

3.7.1

Supported Domains

Pixel Level

3.7.2

Chessboard Settings

Congure the nal chessboard segmentation of the internal thematic layer.

Object Size

Object Size
Level Name

Level name of the level to be created

Reference Book

11 June 2014

Segmentation Algorithms

Overwrite Existing Level

Thematic layer usage ags

3.7.3

Input Parameters

These parameters are identical for the rst and second layers.
Layer

Choose the image layers to analyze. You can disable the second image layer by selecting
No Layer. If you select No Layer, then the other parameters in the second layer group are
inactive.
Scale 14

Several scales can be dened and analyzed at the same time. If at least one scale is
tested positive, the pixels are classied as image objects. The default value is no scale,
indicated by a scale value of 0. To dene a scale, edit the scale value.
The scale value n denes a square (gray) with a side length of 2d (d =
{all pixels with distance to the current pixel |n| 2 + 1 but (|n| 2) 2 + 1)} with
the current pixel in its center. The mean value of the pixels inside the outer square
is compared with the mean value of the pixels inside a inner square (red) with a
side length of 2d (d = {all pixels with distance to the current pixel |n| 2 2 + 1
but not the pixel itself}) . In the case of |n| 3 it is just the pixel value.

Figure 3.10. Scale testing of the contrast filter segmentation

Select a positive scale value to nd objects that are brighter than their surroundings on
the given scale; select a negative scale value to nd darker objects.

Reference Book

11 June 2014

Segmentation Algorithms

Gradient

Use additional minimum gradient criteria for objects. Using gradients can increase the
computing time for the algorithm. Set this parameter to 0 to disable the gradient criterion.
Lower Threshold

Pixels with layer intensity below this threshold are assigned to the Ignored by Threshold
class.
Upper Threshold

Pixels with layer intensity above this threshold are assigned to the Ignored by Threshold
class.

3.7.4

Shape Criteria Settings

If you expect coherent and compact image objects, shape criteria parameters provide an
integrated reshaping operation that modies the shape of image objects by cutting off
protruding parts and lling indentations and hollows.
Shape Criteria Value

Protruding parts of image objects are declassied if a direct line crossing the hollow is
smaller or equal than the ShapeCriteria value. Indentations and hollows of image objects
are classied as the image object if a direct line crossing the hollow is smaller or equal
than the shape criteria value. If you do not want any reshaping, set the value to 0.
Working on Class

Select a class of image objects for reshaping. The pixel classication can be transferred
to the image object level using the classication parameters.

3.7.5

Classification Parameters

Enable Class Assignment

Select Yes or No in order to use or disable the classication parameters. If you select No,
then the other parameters in the group are inactive.
No Objects

Pixels failing to meet the dened lter criteria are assigned the selected class.

Reference Book

11 June 2014

Segmentation Algorithms

Ignored By Threshold

Pixels with a layer intensity below or above the threshold values are assigned the selected
class.
Object in First Layer

Pixels that match the lter criteria in rst layer, but not the second layer, are assigned the
selected class.
Objects in Both Layers

Pixels that match the lter criteria value in both layers are assigned the selected class.
Objects in Second Layer

Pixels that match the scale value in second layer, but not the rst layer, are assigned the
selected class.

Reference Book

11 June 2014

4 Basic Classification
Algorithms
Classication algorithms analyze image objects according to dened criteria and assign
them to a class that best meets them.

4.1

Assign Class

Assign all objects of the domain to the class specied by the Use Class parameter. The
membership value for the assigned class is set to 1 for all objects independent of the class
description. The second and third-best classication results are set to 0.

4.1.1

Use Class

Select the class for the assignment from the drop-down list. You can also create a new
class for the assignment.

4.2

Classification

Evaluates the membership value of an image object against a list of selected classes. The
classication result of the image object is updated according to the class evaluation result.
The three best classes are stored in the image object classication result. Classes without
a class description are assumed to have a membership value of 1.

4.2.1

Active Classes

Erase Old Classification, If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.

Basic Classification Algorithms

Use Class Description

If Yes is selected, class descriptions are evaluated for all classes. The image object
is assigned to the class with the highest membership value.
If No is selected, class descriptions are ignored. This option delivers valuable
results only if Active Classes contains exactly one class
If you do not use the class description, we recommend you use the Assign Class algorithm
instead.

4.3

Hierarchical Classification

The Hierarchical Classication algorithm evaluates the membership value of an image

object against a list of selected classes.
The classication result of the image object is updated according to the class evaluation
result. The three best classes are stored as the image object classication result. Classes
without a class description are assumed to have a membership value of 0. Class-related
features are considered only if explicitly enabled by the according parameter.
This algorithm is optimized for applying complex class hierarchies to entire image object
levels. This reects the classication algorithm of eCognition Professional 4. When
working with domain-specic classication in processes, the algorithms Assign Class
and Classication are recommended.

4.3.1

Active Classes

Choose the list of active classes for the classication.

4.3.2

Use Class-Related Features

Enable to evaluate all class-related features in the class descriptions of the selected classes.
If it is disabled these features will be ignored.

4.4

Remove Classification

Delete classication results from image objects.

4.4.1

Classes

Select classes that should be deleted from image objects.

Reference Book

11 June 2014

5 Advanced Classification
Algorithms
Advanced classication algorithms classify image objects that fulll special criteria, such
as being enclosed by another image object, or being the smallest or largest object in a set.

5.1

Find Domain Extrema

Find Domain Extrema classies image objects with the smallest or largest feature values
within the domain, according to an image object feature.

Figure 5.1. Result of Find Domain Extrema, with Extrema Type set to Maximum and Feature

Set to Area

5.1.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Advanced Classification Algorithms

5.1.2

Extrema Settings

Extrema Type

Choose Minimum to classify image objects with the smallest feature values and Maximum for image objects with the largest feature values.
Feature

Choose the feature to use for nding the extreme values.

Accept Equal Extrema

This feature enables the algorithm to accept equal extrema. This parameter denes the
behavior of the algorithm if more than one image object is fullling the extreme condition.
If enabled, all image objects will be classied; if not, none of the image objects will be
classied.
4.02 Compatibility Mode

Select Yes from the Value eld to enable compatibility with older software versions (version 4.02 and older). This parameter will be removed from future versions.

5.1.3

Classification Settings

Specify the classication that will be applied to all image objects fullling the extreme
condition. At least one class needs to be selected in the active class list for this algorithm.
Active Classes

Choose the list of active classes for the classication.

Erase Old Classification if There is no New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.

Reference Book

11 June 2014

Advanced Classification Algorithms

Use Class Description

5.2

Find Local Extrema

Find Local Extrema classies image objects that fulll a local extrema condition, according to image object features within a search domain in their neighborhoods.
Image objects with either the smallest or the largest feature value within a specic neighborhood will be classied according to the classication settings.

5.2.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

5.2.2

Search Settings

The search settings let you specify a search domain for the neighborhood around the
image object.
Class Filter

Choose the classes to be searched. Image objects will be part of the search domain if
they are classied with one of the classes selected in the class lter. Always add the class
selected for the classication to the search class lter. Otherwise cascades of incorrect
extrema resulting from the reclassication during the execution of the algorithm may
appear.
Search Range

Dene the search range in pixels. All image objects with a distance below the given
search range will be part of the search domain. Use the drop-down arrows to select zero
or positive numbers.
Connected

Enable to ensure that all image objects in the search domain are connected with the analyzed image object via other objects in the search range.

Reference Book

11 June 2014

Advanced Classification Algorithms

3.5 Compatibility Mode

Select Yes from the Value eld to enable compatibility with older software versions (version 3.5 and older). This parameter will be removed in future versions.

5.2.3

Conditions

Dene the extrema conditions.

Extrema Type

Choose Minimum for classifying image objects with the smallest feature values and Maximum for classifying image objects with largest feature values.
Feature

Choose the feature to use for nding the extreme values

Equal Extrema Condition

This parameter denes the behavior of the algorithm if more than one image object is
fullling the extrema condition:
Do Not Accept Equal Extrema means no image objects will be classied
Accept Equal Extrema means all image objects will be classied
Accept First Equal Extrema will result in the classication of the rst of the image
objects.

5.2.4

Classification Settings

Specify the classication that will be applied to all image objects fullling the extreme
condition. At least one class needs to be selected in the active class list for this algorithm.
Active Classes

Choose the list of active classes for the classication.

Erase Old Classification if There is no New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.

Reference Book

11 June 2014

Advanced Classification Algorithms

Use Class Description

5.3

Find Enclosed by Class

Find and classify image objects that are completely enclosed by image objects belonging
to certain classes.
If an image object is located at the border of the image, it will not be found and classied
by Find Enclosed by Class. The shared part of the outline with the image border will not
be recognized as the enclosing border.

Figure 5.2. Left: Input of Find Enclosed by Class: Domain: Image Object Level, Class Filter:

N0, N1. Enclosing class: N2. Right: Result of Find Enclosed by Class: Enclosed objects get
classified with the class enclosed. Note that the objects at the upper image border are not
classified as enclosed

5.3.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

5.3.2

Search Settings

Enclosing Classes

Choose the classes that might be enclosing the image objects

Reference Book

11 June 2014

Advanced Classification Algorithms

Compatibility Mode

Select Yes from the Value eld to enable compatibility with older software versions (version 3.5 and 4.0). This parameter will be removed with future versions.

5.3.3

Classification Settings

Choose the classes that should be used to classify enclosed image objects.
Active Classes

Choose the list of active classes for the classication.

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

5.4

Find Enclosed by Image Object

Find and classify image objects that are completely enclosed by image objects from the
domain. Enclosed image objects located at the image border will be found and classied
by the Find Enclosed by Image Object algorithm. The shared part of the outline with the
image border will be recognized as the enclosing border.
In gure 5.3 on the following page, the inputs for the left-hand image in Find Enclosed by
Image Object are domain: image object level, class lter: N2. On the right-hand image,
using the same algorithm, enclosed objects are classied with the class enclosed. Note
that the objects at the upper image border are classied as enclosed.

5.4.1

Classification Settings

Choose the class that will be used to classify enclosed image objects.

Reference Book

11 June 2014

Advanced Classification Algorithms

Figure 5.3. Comparison of Find Enclosed by Class

Active Classes

Choose the list of active classes for the classication.

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

5.5

Connector

Classify the image objects that make up the shortest connection between the current image object and another image object that meets the conditions described by the connection
settings.
The process starts to search from the current image object along objects that meet the
conditions as specied by Connect Via and Super Object Mode Via, until it reaches an

Reference Book

11 June 2014

Advanced Classification Algorithms

image object that meets the conditions specied by Connect To and Super Object Mode
To. All image objects that are part of the resulting connection are assigned to the selected
class.
You can dene the maximum search range under Search Range In Pixels.

5.5.1

Connector Settings

Connector Via

Choose the classes whose image objects you want to be searched when determining the
connection. Image objects assigned to other classes will not be taken into account.
Super Object Mode Via

Limit the number of image objects that are taken into account for the connection by
specifying a superordinate object. Choose one of the following options:
Dont Care: Use any image object
Different Super Object: Use only images with a different superobject than the seed
object
Same Super Object: Use only image objects with the same superobject as the seed
object
Connect To

Choose the classes whose image objects you want to be searched when determining a
destination for the connection. Image objects assigned to other classes will not be taken
into account.
Super Object Mode To

Limit the number of image objects that are taken into account for the destination of the
connection by specifying a superordinate object. Choose one of the following options:
Dont Care use any image object
Different Super Object use only images with a different superobject than the seed
object
Same Super Object use only image objects with the same superobject as the seed
object.
Search Range

Enter the Search Range in pixels that you wish to search.

Reference Book

11 June 2014

Advanced Classification Algorithms

5.5.2

Classification Settings

Choose the class that should be used to classify the connecting objects.
Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

5.6
5.6.1

Assign Class By Slice Overlap (Prototype)

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

5.6.2

Algorithm Parameters

Use Class

Select class for assignment

Candidate Domain

Class lter for objects that are considered.

Create 3D Objects

Create 3D objects

Reference Book

11 June 2014

Advanced Classification Algorithms

Slices Up

Describes how many slices the object is copied up. Use 1 for no limit.
Slices Down

Describes how many slices the object is copied down. Use 1 for no limit.

5.7

Optimal Box (Prototype)

Generate member functions for classes by looking for the best separating features based
upon sample training.

5.7.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

5.7.2

Sample Class Parameters

For Target Samples

Select a class or create a new class that provides samples for the target class (the class to
be trained).
For Rest Samples

Select a class or create a new class that provides samples for the rest of the domain.

5.7.3

Insert Membership Function Parameters

For Target Samples Into

Select a class or create a new class that receives membership functions after optimization
for target. If set to unclassied, the target sample class is used.
For Rest Samples Into

Select a class or create a new class that receives inverted similarity membership functions
after optimization for target. If set to unclassied, the rest sample class is used.

Reference Book

11 June 2014

Advanced Classification Algorithms

Clear All Membership Functions

When inserting new membership functions into the active class, choose whether to clear
all existing membership functions or clear only those from input feature space:
No, Only Clear if Associated with Input Feature Space: Clear membership functions only from the input feature space when inserting new membership functions
into the active class.
Yes, always clear all membership functions: Clear all membership functions when
inserting new membership functions into the active class.

5.7.4

Feature Optimization Parameters

Input Feature Set

Input set of descriptors from which a subset will be chosen. Click the ellipsis button
to open the Select Multiple Features dialog box. The Ensure Selected Features are in
Standard Nearest Neighbor Feature Space checkbox is selected by default.
Minimum Number of Features

The minimum number of features descriptors to employ in a class. The default is 1.

Maximum Number of Features

The minimum number of features descriptors to employ in a class.

5.7.5

Optimization Settings Parameters

Weighted Distance Exponent

Enter a number greater than 0 to decrease weighting with increasing distance. Enter 0 to
weight all distances equally. The default is 2.
Simplification Factor Mode

The simplication factor is calculated as:

cur dimension
total dim

The following values are available:

Do Not Apply: Simplication factor is not applied
Apply to Weighted Distance: Weighted distance is multiplied by simplication
factor trades off simplicity against distance separation for equally good overlap
Apply to Overlap: Overlap is divided by simplication factor trades off simplicity
against false positives.

Reference Book

11 June 2014

Advanced Classification Algorithms

Simplification Factor Exponent

The exponent where the simplication factor is calculated as:

cur dimension
total dim

5.7.6

Optimization Output Parameters

False Positives Variable

Set this variable to the number of false positives after execution.

False Negatives Variable

Set this variable to the number of false negatives after execution.

False Positives Object List

Select an existing image object list variable, to which image objects associated with false
positive samples are added. If a value is not set, no output occurs. Please note that
samples stored in the workspace have no associated objects, so cannot be included in the
list.
False Negatives Object List

Select an existing image object list variable, to which image objects associated with false
negative samples are added. If a value is not set, no output occurs. Please note that
samples stored in the workspace have no associated objects, so cannot be included in the
list.
Separated Positives Object List (For Selected Feature Space)

This list contains image objects associated with positive samples, which are separated
from other positive samples by false positives, using the generated feature space.
Separated Positives Object List (From For Target Samples Into Class)

This list contains image objects associated with positive samples, which are separated
from other positive samples by false positives, using a feature space from For Target
Samples Into.
Show Info in Message Console

Show information on feature evaluations in message console.

Reference Book

11 June 2014

Advanced Classification Algorithms

5.8

Classifier

The Classier algorithm lets you apply machine-learning functions to your analysis in
a two-step process: First a classier is trained using the classied objects of the domain as training samples. The trained classier is stored as a string variable in the
conguration settings. In a second step, the trained classier from the rst step is applied to the domain, classifying the image objects according to the trained parameters.
For a more detailed description and example project of selected classiers please refer
to our user community: http://community.ecognition.com/home/CART%20-%20SVM
%20Classier%20Example.zip/view
The Classier algorithm can be applied either object based or image layer based (pixel
based).

Figure 5.4. The training step

Figure 5.5. The application step

5.8.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Advanced Classification Algorithms

5.8.2

General Parameters

Operation

Choose the operation to perform:

Train: Train a classier. The domain objects serve as training samples.
Apply: Apply a trained classier to the domain objects.
Query: Query attributes of a trained classier.

5.8.3

Train: General Parameters

Configuration

You can enter or select a string variable to store or load the classier conguration.
Use Samples Only

Choose Yes or No depending on whether you want to use sample image objects only.

5.8.4

Train: Feature Space

Source - object based

Uses image objects as input source with the following options:

Type Specify whether to use object features or a feature array.
object features
* Features: Select object features for classication.
feature array
* Feature array: Select an array of features for classication
Source - layer based

Uses pixel based layers as input source:

Layer Array: Array of layers to use as pixel based features.

5.8.5

Train: Classifier Parameters

Select a classier to train from Decision Tree, Random Trees, Bayes, KNN or SVM.
The following options are displayed selecting the Algorithm classier - Parameter Operation Train - Type:
KNN
Features: Select the features for classication.
Normalize: Change to Yes to normalize the selected features(No default).

Reference Book

11 June 2014

Advanced Classification Algorithms

k: Insert the value for k.

SVM
Features: Select the features for classication.
Normalize: Change to Yes to normalize the selected features(No default).
Kernel type: Select either linear or RBF for kernel type (linear default).
C: Insert the SVM parameter C (2 default).
gamma: For kernel type RBF insert the SVM parameter gamma (0 default).
Decision Tree
Depth: Maximum tree depth (0 default).
Min sample count: Minimum number of samples per node (0 default).
Use surrogates: Use surrogates for missing data. If yes then surrogate splits
will be built to be able to work with missing data.
Max categories: Cluster possible values of a categorical variable into K <
max_categories clusters (16 default).
Cross validation folds: Number of cross validations to perform (3 default).
Use 1 SE rule: Select Yes to use the 1SE rule for pruning (No default).
Truncate pruned tree: Pruned branches are physically removed from the
tree if set to Yes (Yes default).

Random Trees
Depth: Maximum tree depth (0 default).
Min sample count: Minimum number of samples per node (0 default).
Use surrogates: Use surrogates for missing data. If yes then surrogate splits
will be built to be able to work with missing data.
Max categories: Cluster possible values of a categorical variable into K <
max_categories clusters (16 default).
Active variables: The size of the randomly selected subset of features at each
tree node and that are used to nd the best split(s). (0 default - if you set it
to 0 then the size will be set to the square root of the total number of features).
Max tree number: The maximum number of trees (50 default).
Forest accuracy: Sufcient accuracy of trained forest in % (0.01 default).
Termination criteria type: The type of learning termination criteria allows
to decide how training should be stopped: by max number of trees or forest
accuracy or both of them (both default).

5.8.6

Apply: General Parameters

Configuration

Select the string variable to load and apply your classier conguration.

5.8.7

Apply: Feature Space

Apply the settings in accordance with the training parameters.

Reference Book

11 June 2014

Advanced Classification Algorithms

Source - object based

Uses image objects as input source.

Source - layer based

Uses pixel based layers as input source:

Layer Array: Array of layers to use as pixel based features.

5.8.8

Query: General Parameters

Configuration

Select the string variable to load the respective classier conguration.

5.8.9

Query: Query Information about a trained classifier

Type

Select from:
Bayes
KNN
SVM
The following options are displayed selecting the Algorithm classier - Parameter Operation Train - Type:
Decision Tree
Operation. Select from:
* Query Parameter:
Depth
Min Sample Count
Cross validation folds
Max categories
Importance Feature Array: Creates an array that contains importance features based on a trained classier.
Importance Value Array: Creates a double array that contains only
importance values within the importance feature array.
Export Importance Table: Exports all features of the input feature
space with corresponding importance values.
Query
Node:
*
ld: Specify the tree node to query properties from.
ld leaf: Specify the variable to receive a property.
Assigned Class: Specify the class variable to receive a property.
Left Child Node: Specify the variable to receive a property.
Right Child Node: Specify the variable to receive a property.
Split Feature Index: Specify the variable to receive a property.
Split Threshold: Specify the variable to receive a property.

Reference Book

11 June 2014

Advanced Classification Algorithms

* Plot: Export a diagram of the decision tree as a Windows bitmap.

Image Path: Enter the image path to export the image le.
* Convert to CNL: Generate CNL code, that is equivalent to the decision
tree.
Target Process Path: Location of the CNL tree.
Prefix: Add a prex.
Random Trees
Depth: Maximum tree depth.
Min Sample Count: Minimum number of samples per node.
Max categories: Cluster possible values of a categorical variable into K <
max_categories clusters .
Active variables: The size of the randomly selected subset of features at each
tree node and that are used to nd the best split(s).
Max tree number: The maximum number of trees.
Forest accuracy: The accuracy of trained forest in %.
Actual tree number: The number of trees generated.
Importance Feature Array: Creates an array that contains importance features based on a trained classier.
Importance Value Array: Creates a double array that contains only importance values within the importance feature array.
Export Importance Table: Exports all features of the input feature space
with corresponding importance values.
For more information about classier functions, see the OpenCV C++ Reference. 1

1. docs.opencv.org/modules/ml/doc/ml.html

Reference Book

11 June 2014

6 Template Matching
High quality templates are critical for achieving optimal results with the Template Matching algorithm (p 162). To facilitate the generation of such templates, eCognition offers
a sophisticated template editor. In this editor the user can conveniently select and adjust
sample representations of the template in the image. A fully automatic algorithm is then
applied to extract generic templates that capture what the samples have in common. It
is further possible to test the generated templates on a subregion, giving the user a quick
overview over the number and quality of targets detected. Targets that the user identies
as correct are automatically added as samples, thus increasing the sample number and
allowing for iterative renement of template quality. To open the dialog select View >
Windows > Template Editor. The Template Matching algorithm (p 162) can be applied
in the Process Tree window.
See also the template matching demo projects provided in the eCognition community
http://www.ecognition.com/community.

6.1

Overview Template Editor

To create a new template open the Template Editor and click the folder button.

You can create several templates for a given project if there are different kinds of targets
you would like to nd on a given image. The generated templates are stored with your
project when you save it. Note that activating the Select sample mode without an existing
template will open the Create New Template dialog automatically.
To select the root folder to store the templates that you generate, select the Browse button.
Default location is the root folder of your workspace directory, if you work with projects,
your templates are stored in the image folder by default. It is generally not necessary
to change this folder location later on, but if you do, you may need to regenerate your
templates using the Generate Templates tab.
To delete a template including its samples select the respective button:
To select the zoom size of selected samples, templates or targets use the Thumbnail zoom
in the respective tab.

Template Matching

Figure 6.1. The Template Editor Dialog

6.2

Select Samples

Use the Select samples tab to select examples of your template in the image. Selected
samples are stored within the project when you save your project.
To select samples you must activate the Select Samples button. The mouse pointer
changes to precision selection mode and the samples can be inserted by a single click
into the image. The extent of the sample used for template calculation can be inserted
in the eld Size (in pixels). A red rectangle of the selected sample size is drawn around
the samples center point. You can also adapt the sample size after the selection of samples. The eld Context lets you adjust the size of the displayed image context of your
thumbnail preview for better visual evaluation. The context is not considered in the template calculation. The image layer mixing of the thumbnail preview corresponds with the
settings of the main view.
To delete samples right click and select Delete Sample(s) either in the main view (single
selection only) or in the thumbnail preview window of the Template Editor (multiple
selection possible).
To adjust the exact position of the sample crosshair for selected samples:
1. In the Template Editor dialog activate the button Adjust Center and click on a
thumbnail to obtain an enlarged visualization of the sample. Now you can shift the
crosshair.
2. In the main view move the mouse over the crosshair so that it changes from red to
black. (The select samples button must be active.)
To obtain convincing results you should try to capture the variability that the template
Reference Book

11 June 2014

Template Matching

Figure 6.2. Selecting Samples

displays in actual images. If variability cannot be addressed in a single image, the user is
advised to create a multi-scene project and collect samples from different scenes. Note
that it is important for template generation that samples are all centered on the same
corresponding location.

6.3

Generate Templates

This tab allows you to generate template images. You can select the layer for which
the template is generated, and a group size parameter. When the group size is set to 0
(default value), the generated template reects the mean of all samples. For group size >
0 the algorithm attempts to nd subgroups in the samples and creates a separate template
for each subgroup. To generate templates click on the Generate Template button. All

Reference Book

11 June 2014

Template Matching

Figure 6.3. Adjusting the Center of selected Samples

templates generated are stored separately according to the image layer name selected and
the group size chosen e.g. \New Template\Layer2_Group5. (If you want to visualize the
templates properly outside eCognition please select an image viewer that supports 32-bit
oating image les.)
The generated templates are also displayed in the thumbnail view of the Template Matching dialog. The average correlation coefcient between the template and all samples is
displayed at the bottom of the window. For template groups the maximum correlation
between this sample and any of the templates are considered for each sample before averaging across samples.

6.4

Test Template

In this tab you can test a given template on a subregion and identify the targets that are
detected with a user-dened threshold. You can then assign the targets as correct or false
to obtain an estimate of error rates.

Reference Book

11 June 2014

Template Matching

Figure 6.4. Generate Templates - Group Size 5

6.4.1

Subset Selection

To select a subset for testing activate the Select Region button and then either draw a
rectangle in the main view or insert coordinates. To obtain meaningful results for the
Template Quality select a region that includes most of your samples (see Template Quality
on page 72).

6.4.2

Template Selection

Select a layer and group size used when generating the template.

6.4.3

Test

To start the template matching algorithm, select Execute Test. If Update template is
activated, the current template is regenerated based on the current samples before the test
is executed. This means that targets that have been classied as correct in a previous
test run are reected in the template as well.

6.4.4

Test Parameters

By changing the values for the Ground Truth tolerance (in pixels) you can dene
how accurately a found candidate must match a sample to be identied as correct.
A higher value means more candidates will be found.
Reference Book

11 June 2014

Template Matching

Figure 6.5. Test Template

Threshold r for candidates. (Trimble recommends starting with the sample correlation value of the Generate Samples tab on page 68. If you insert low values the
computation time increases as more candidates will be considered.)

6.4.5

Review Targets

After executing, you can assign the test candidates as Correct, False or Not Sure using the
respective buttons. The results are shown in the three small windows below: Unclassied,
Correct and False. Targets that are classied as correct are automatically added to the
samples (see Select Samples tab on page 67). The assignments are stored with the project
after saving and do not have to be repeated in subsequent tests.

Reference Book

11 June 2014

Template Matching

6.4.6

Template Quality

This section of the dialog provides information on the number of correctly, false, unclassied and missed targets and the corresponding false alarm and missed target rate in
percent. Note: The missed target rate is accurate only if you have identied all possible samples in your test region. (Otherwise, the software cannot know if targets were
missed.)

6.5

Negative Templates

In this tab, you can create negative templates. Negative templates allow you to improve
your results by eliminating false positives.
Before creating a negative template, the user should optimize the actual positive (normal) template. The nal test with the positive template should be done on a large subregion, using a relatively low threshold, so that many false positives are obtained. Also
make sure that all unclassied items that do not reect correct targets are indeed classied
as false.
In the Negative Templates tab click on the Generate Templates button to generate a negative template or template group. The template is generated according to the same algorithm as the positive template, but instead of using the selected samples as a base, it
uses the false positives of the last test as a base. You can create negative templates with
different group sizes, but you cannot change the layer, as the layer always corresponds to
the layer of the current (positive) template. The negative templates are stored inside the
folder of the corresponding positive template.
The functionality is otherwise identical to the Generate Samples tab on page 68.

Reference Book

11 June 2014

7 Variables Operation
Algorithm
Variable operation algorithms are used to modify the values of variables. They provide
different methods for performing computations based on existing variables and image
object features, and for storing the results within variables.

7.1

Timer

The Timer algorithm adds to a variable the elapsed time of its sub-processes (with the exception of customized algorithms). It is useful for debugging and improving performance
in rule sets.

7.1.1

Supported Domains

Execute

7.1.2

Algorithm Parameters

Timer Variable

Select the variable to which the process execution time and all its subprocesses will be
added.

7.2

Calculate Random Number

This algorithm calculates a random value and stores it in a variable. To initialize (seed)
the randomization set this variable to 0 (outside a loop).

Variables Operation Algorithm

7.2.1

Algorithm Parameters

Variable for the random value

Variable to store the computed random value. Please initialize this randomization algorithm by setting this variable to 0.

7.3

Update Variable

Perform an arithmetic operation on a variable.

7.3.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List; Array

7.3.2

Algorithm Parameters

Variable Type

Object, scene, feature, class, level, image layer, thematic layer, or map variables must be
selected as the vriable type. Select the variable assignment, according to the variable type
selected in the Variable Type eld. To select a variable assignment, click in the eld and
do one of the following depending on the variable type:
For object variables, use the drop-down arrow to open the Select Single Feature
dialog box and select a feature or create a new feature variable
For scene variables, use the drop down arrow to open the Select Single Feature
dialog box and select a feature or create a new feature variable
For feature variables, use the ellipsis button to open the Select Single Feature dialog
box and select a feature or create a new feature variable
For class variables, use the drop-down arrow to select from existing classes or
create a new class
For level variables, use the drop-arrow to select from existing levels
For image layer variables, select the image layer you want to use for the update
operation
For thematic layer variables, select the thematic layer you want to use for the update
operation
For map variables, select the map name you want to use for the update operation
For array variables, select the array you want to use for the update operation.
Variable

Select an existing variable or enter a name to create a new one. If you have not already
created a variable, the Create Variable dialog box will open. Only scene variables can be
used in this eld.

Reference Book

11 June 2014

Variables Operation Algorithm

Operation

This eld displays only for object and scene variables. Select one of the arithmetic operations:
Value Description
=

Assign a value.

Increase by value.

Decrease by value.

Multiply by value.

Divide by value.

Assignment

If Scene Variable or Object Variable is selected, you can assign either a value or a feature.
This setting enables or disables the remaining parameters. If Image Layer Variable or
Thematic Layer Variable is selected, you can assign either a layer or index.
Value

This eld displays only for Scene and Object variables. If you have selected to assign by
value, you may enter a value or a variable. To enter text use quotes. The numeric value
of the eld or the selected variable will be used for the update operation.
Feature

This eld displays only for scene and object variables. If you have chosen to assign by
feature you can select a single feature. The feature value of the current image object will
be used for the update operation.
Comparison Unit

This eld displays only for Scene and Object variables. If you have chosen to assign
by feature, and the selected feature has units, then you may select the unit used by the
process. If the feature has coordinates, select Coordinates to provide the position of the
object within the original image or select Pixels to provide the position of the object
within the currently used scene.
Arithmetic Expression

For all variables, you may assign an arithmetic expression to calculate a value.

Reference Book

11 June 2014

Variables Operation Algorithm

Array Item

The Array Item parameter appears when by array item is selected in the Assignment
parameter. You may create a new array item or assign a Domain.

7.4

Compute Statistical Value

Perform a statistical operation on the feature distribution within a domain and store the
result in a scene variable.

7.4.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

7.4.2

Active Classes

Choose the list of active classes for the classication.

7.4.3

Algorithm Parameters

Variable

Select an existing variable or enter a name to create a new one. If you have not already
created a variable, the Create Variable dialog box will open. Only scene variables can be
used in this eld.
Operation

Select one of the statistical operations listed in table 7.0 on the next page, Available
Operations for Compute Statistical Value Algorithm.
Parameter

If you have selected the quantile operation, specify the percentage threshold.
Feature

Select the feature that is used to perform the statistical operation. (This parameter is not
used if you select Number as your operation.)

Reference Book

11 June 2014

Variables Operation Algorithm

Table 7.0. Available Operations for Compute Statistical Value Algorithm

Value

Description

Number

Count the objects of the currently selected domain

Sum

Return the sum of the feature values from all objects of the selected
domain

Maximum

Return the maximum feature value from all objects of the selected domain

Minimum

Return the minimum feature value from all objects of the selected domain

Mean

Return the mean feature value of all objects from the selected domain

Standard
Deviation

Return the standard deviation of the feature value from all objects of the
selected domain

Median

Return the median feature value from all objects of the selected domain

Quantile

Return the feature value, where a specied percentage of objects from the
selected domain have a smaller feature value

Unit

You may select a unit for the operation

For positional features (such as X Max) you can choose Pixels from the drop-down
list to return the relative object position (such as the position within a tile or subset).
If you select a positional feature you can select Co-ordinates in the drop-down list
to use the absolute co-ordinates if desired.

7.5

Compose Text

Assign a string variable text parts. Content of variable will be replaced by the text assigned with parameters text prex and text sufx.

7.5.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

7.5.2

Algorithm Parameters

Result

Enter the name of the variable that will store the composed text.

Reference Book

11 June 2014

Variables Operation Algorithm

Text Prefix

Edit the beginning of the composed text: the text sufx is attached afterwards. You can
enter one of the following:
Text (between quotation marks)
A number value (without quotation marks)
A variable (without quotation marks)
Text Suffix

Edit the end of the composed text: it is attached after the text prex. You can enter one
of the following:
Text (between quotation marks)
A number value (without quotation marks)
A variable (without quotation marks)
Examples

Some possible combinations are shown below:

Text prefix

Type

Text suffix Type

class

Text

CLASSNAMEID Variable ID

7.6

Composed Text

Number value class1

Variable

class1

Update Region

Modify a region variable. You can resize or move the region dened by a region variable,
or enter its coordinates. Alternatively, you can use the coordinates of an image object
bounding box or the active pixel for a region update.

7.6.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Array

7.6.2

Algorithm Parameters

Variable

Select an existing region variable or create a new one. To create a new variable, type a
name for the new variable and click OK to open the Create Region Variable dialog box
for further settings.

Reference Book

11 June 2014

Variables Operation Algorithm

Mode

Select the operation for modifying the region dened by the region variable (see table 7.1,
Update Region Algorithm Modes).
Table 7.1. Update Region Algorithm Modes

Value

Description

Set min/max
coordinates

Set each existing coordinate of the region by entering values for minimum
and maximum coordinates, see below.

Set by
origin/extent

Set each existing coordinate of the region by entering values for the Origin
and the Extent, see below.

Move

Move a region by entering absolute Move in values for each existing

coordinate, see below.

Resize

Shrinking or grow a region from its center. Enter absolute or percent

values for each existing coordinate, see Unit and Resize in below. Because
growing is measured from the center, the resulting region co-ordinates
might be negative. In this case, the region is shifted so that the respective
origin coordinate is 0. Examples:
If a region (100, 100), [100, 100] is grown by 100% the result is
(50, 50), [200, 200]; no shifting is needed.
If a region (0, 0), [100, 100] is grown by 100% the result would be
(50, 50), [150, 150]; instead the region is shifted by +50 in x- and y
direction to (0, 0), [200, 200].

From object

Use coordinates of the bounding box including all image objects in the
domain.

From array

Select a user-dened array (for more information, please consult the user
guide).

Check bounds

A region can be fully or partly outside the scene, for example after
initializing a region variable from a main map to use it in a rescaled map.
This mode makes sure that region is tted within the scene specied in the
process domain.Examples:
If a region (100, 100), [9999, 9999] should be applied to a scene of
(500,500), select Check bounds to truncate the region to (100, 100),
[500, 500]
If a region (100, 100),[9999, 9999] should be applied to a scene of
(500,500), select Check bounds to truncate the region to (0, 0),
[500, 500].

Active Pixel

Dene a region based on the active pixel/voxel. The co-ordinates of the

active pixel/voxel are used for the origin of a region with the extent of 1 in
each dimension.

Assign

Set the region to the same values like the region specied in Assign
Region.

Reference Book

11 June 2014

Variables Operation Algorithm

7.7

Update Image Object List

The Update Image Object List algorithm allows you to modify list variables, add objects,
remove objects, and clear and compact lists.

7.7.1

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Superobject; Sub-objects; Linked Objects

7.7.2

Algorithm Parameters

Variable

Select an existing image object list variable or create a new one. The computed value will
be stored in the image object list variable.
Mode

Select the operation to use to modify the object list.

Add Image Objects adds all objects in the current domain to the selected image
object list. All image object domains are supported.
Remove Image Objects removes all image objects in the current domain from the
list. All image object domains are supported.
Clear List removes all image objects from the selected image object list. The Execute domain is supported.
Compact List removes all invalid objects from the list. The Execute domain is
supported.
Feature

Select which feature you wish to use for sorting.

Sorting Order

Select the sorting order for the image objects.

7.8

Update Feature List

Update Feature List lets you to create a list that contains features, which can be later
exported (for example as project statistics). The feature list can contain any combination
of features, all of which can be added and removed from a list. Features can also be added
and removed between lists and entire lists can be deleted.

Reference Book

11 June 2014

Variables Operation Algorithm

7.8.1

Supported Domains

Execute

7.9

Automatic Threshold

Calculates a threshold value to be used for multi-threshold segmentation.

Typical usage of automatic threshold algorithm in combination with multithreshold segmentation

Figure 7.1.

Figure 7.2. Result of multi-threshold segmentation with rule set as described above

The threshold can either be determined for an entire scene or for individual image objects.
Depending on this setting, it is stored in a scene variable or in an image object variable.
You can use the variables as a parameter for multi-threshold segmentation.

Figure 7.3. Use of variables as parameters in a multi-threshold segmentation

Using the automatic threshold algorithm in combination with multi-threshold segmentation allows you to create fully automated and adaptive image analysis algorithms based
on threshold segmentation. A manual denition of xed thresholds is not necessary.
Reference Book

11 June 2014

Variables Operation Algorithm

The algorithm uses a combination of histogram-based methods and the homogeneity measurement of multi-resolution segmentation to calculate a threshold dividing the selected
set of pixels into two subsets, so that heterogeneity is increased to a maximum.

7.9.1

Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

7.9.2

Algorithm Parameters

Image Layer

Select the image layer you want to be used for automatic threshold determination.
Value Range

Dene the value range that is taken into account for automatic threshold determination.
Entire Value Range the complete threshold value range is taken into account.
Restricted Value Range the threshold values within the specied interval (Min.
Value Max.Value) are taken into account.
Min. Value

If the value range parameter is set to Restricted Value Range, dene a minimum value for
the automatically determined threshold.
Max. Value

If the value range parameter is set to Restricted Value Range, dene a maximum value
for the automatically determined threshold.
Threshold

Select the variable where the threshold value is to be saved. You can either select a numeric scene variable or a numeric image object variable (type Double). Saving threshold
values in scene variables is most useful in combination with the Pixel Level domain. Saving threshold values in image object variables allows you to differentiate between the
image objects of different domains.

Reference Book

11 June 2014

Variables Operation Algorithm

Quality

Select the variable in which a quality control parameter for the threshold value is to be
stored. The quality control parameter reects the internal criterion according to which
the threshold is determined. The higher its value, the more distinct the segmentation
into image objects. You can use the quality control parameter to implement fully automated segmentation processes splitting up the domain into image objects that meet the
predened requirements.

7.10 Update Array

The Update Array algorithm is used to update user-dened arrays in rule sets. For more
information on creating arrays, please consult the eCognition Developer 9.0 User Guide
> Advanced Rule Set Concepts > Arrays and have a look at the examples in our user
community:
http://community.ecognition.com/home/Arrays%20Example%20%231.zip/view
or
http://community.ecognition.com/home/Arrays%20Example%20%232.zip/view

7.10.1 Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

7.10.2 Algorithm Parameters

Array

Select an array from the drop-down list.

Mode

The following options are available to modify the array:

1.
2.
3.
4.
5.
6.
7.
8.

Add Values: Select a value to add to the array

Add Based From Array Items: Add a value from a specied array
Add Based From Feature: Add a value to an array derived from a feature
Update Value: Enter an index or value pair for the value to be updated
Remove By Value: Select a value to remove from the array
Remove By Index: Select the index of the value to be removed
Clear: Remove all values from the array
Sort: Sort array values in ascending or descending order.

Selecting options 13 will display the Add Value parameter. Select only once for a
single value; selecting always allows duplicates.

Reference Book

11 June 2014

8 Basic Object Reshaping

Algorithms
Basic reshaping algorithms modify the shape of existing image objects. They execute
operations such as merging image objects and splitting them into their sub-objects.

8.1

Remove Objects

Merge image objects in the image object domain. Each image object is merged into
the neighbor image object with the largest common border. This algorithm is especially
helpful for clutter removal.

8.1.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

8.1.2

Algorithm Parameters

Target Class

Click the ellipsis button to open the Edit Classication Filter dialog box and select the
target classes to which the image objects in the image object domain will be merged.
Show Advanced Parameters

Changing this value to Yes activates the following options:

Merge By allows you merge image objects based on shape or color. Merging
by shape merges image objects to their neighbors, based on the longest common border.
Merging by color merges neighbors based on the smallest layer value difference.
Merge By

Merging by shape activates the following parameters:

If Use Threshold has a value of Yes, the Border Threshold parameter is activated

Basic Object Reshaping Algorithms

Enter a value in Border Threshold only image objects with a common border
length longer than or equal to this threshold will be merged
If Use Legacy Mode is set to Yes, the algorithm will not look for a common
superobject. (If a superobject level exists, objects may not be completely
removed.)
Merging by color activates the following parameters:
If Use Threshold has a value of Yes, the Color Threshold parameter is activated.
In Layer Usage, select the layers to be analyzed
Enter a value in Color Threshold only image objects with a color difference
smaller than or equal to neigboring image objects will be merged. For all image objects in the domain the sum of absolute mean layer differences over all
selected layers is calculated SUM_ALL_LAYERS(ABS(MeanColor(object)
- MeanColor(neighbour))). An object is merged with the neighboring object
where the resulting value is the smallest (minimum color difference).
If Use Legacy Mode is set to Yes, the algorithm will not look for a common
superobject. (If a superobject level exists, objects may not be completely
removed.)

8.2

Merge Region

Merge all image objects in the image object domain.

Figure 8.1. Result of merge region algorithm on all image objects classified as Seed

8.2.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Basic Object Reshaping Algorithms

8.2.2

Algorithm Parameters

Fusion Super Objects

Enable the fusion of afliated super objects.

Use Thematic Layers

Enable to keep borders dened by thematic layers that were active during the initial segmentation of this image object level.

8.3

Grow Region

Enlarge image objects dened in the image object domain by merging them with neighboring image objects (candidates) that match the criteria specied in the parameters.
The grow region algorithm works in sweeps. That means each execution of the algorithm
merges all direct neighboring image objects according to the parameters. To grow image
objects into a larger space, you may use the Loop While Something Changes checkbox
or specify a specic number of cycles.

Figure 8.2. Result of grow region algorithm on image objects of class Seed and candidate class

not Vegetation.

8.3.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Basic Object Reshaping Algorithms

8.3.2

Algorithm Parameters

Candidate Classes

Choose the classes of image objects that can be candidates for growing the image object.
Fusion Super Objects

Enable the fusion of afliated super objects.

Candidate Condition

Choose an optional feature to dene a condition that neighboring image objects need to
fulll in addition to be merged into the current image object.
Use Thematic Layers

Enable to keep borders dened by thematic layers that were active during the initial segmentation of this image object level.

8.4

Convert to Sub-objects

Split all image objects of the image object domain into their sub-objects.

8.4.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

8.4.2

Algorithm Parameters

None

8.5

Convert Image Objects

Convert image objects to a specied type, based on the differences among image object
types with regard to their spatial connectivity and dimensions.

8.5.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects

Reference Book

11 June 2014

Basic Object Reshaping Algorithms

8.5.2

Algorithm Parameters

Image Object Type

Convert all image objects in the image object domain to image objects of the specied
type:
Type

Description

Connected 2D Convert all image objects in the image object domain to connected 2D image
objects. The conversion is performed only for image objects that are not of the
type connected 2D.
Connected 3D Convert all image objects in the image object domain to connected 3D image
objects. Internally, all image objects in the image object domain are converted
to connected 2D image objects. Connected 3D image objects are then created
based on the overlaps of connected 2D image objects over the slices.
Connected 3D image objects are continuous with respect to slices. They may
have several disconnected parts within a single slice.
A special situation occurs if you have multiple image object levels and some
parts of a connected 3D image object belong to different superobjects. In that
case, the superobjects are merged automatically. If merging is not possible,
then single disconnected superobjects are generated.
Disconnected

Convert all image objects in the image object domain to disconnected image
objects. The algorithm tries to create a single image object per class. If some
3D image objects to be merged belong to different superobjects, the conversion
works according to the Fusion of Superobjects settings; see below.

Figure 8.3. Merging scheme of overlapping image objects between slices.

Reference Book

11 June 2014

Basic Object Reshaping Algorithms

Fusion of Super Objects

This parameter enables you to specify the effect on super objects. If this parameter is set
to Yes, super objects will also be fused. This will usually have the effect that affected
super objects will be converted to disconnected objects. If this parameter is set to No, the
super objects stay untouched. In this case, fusion of the active image objects is restricted
by the extent of the super objects.
If the value is yes, superobjects are merged. If superobjects cannot be merged into
a single connected image object they are merged into single disconnected image
objects (default)
If the value is no, superobjects are not merged and only image objects having the
same superobject are merged. Consequently, there can be several disconnected
image objects per class.

8.6

Cut Objects at Region

Cut all image objects within the image object domain that overlap the border of a given
region. Each image object to be cut is split into two image objects, called pieces: one is
completely within the region and one is completely outside.

8.6.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects

8.6.2

Algorithm Parameters

Region

Select or enter the name of an existing region. Alternatively, you can enter the coordinates of a region specied by its origin (xG , yG ) , which is the lower left corner, and
its size [Rx , Ry , Rz , Rt ]. The input pattern is: (xG , yG , zG ,tG ), [Rx , Ry , Rz , Rt ]. Alternatively,
you can select a variable. To create a new variable, type a name for the new variable and
click OK or press Enter to open the Create Variable dialog box for further settings.
Object Type Inner Pieces

Dene the type of all image objects (pieces) located inside the region, no matter if they
are cut or not:

Reference Book

11 June 2014

Basic Object Reshaping Algorithms

Value

Description

Keep current

Keep current type (default)

Connected 2D Convert resulting image objects to connected 2D image objects.

Connected 3D Convert resulting image objects to connected 3D image objects (only available
with eCognition Developer 9.0)
Disconnected

Convert resulting image objects to disconnected image objects.

Classify Inner Pieces

Select Yes to classify all image objects (pieces) located inside the region, no matter if
they are cut or not.
Object Type Outer Pieces

Dene the type of all cut image objects (pieces) located outside the region. The same
settings are available for Object Type Inner Pieces.
Classify Outer Pieces

Select Yes to classify all cut image objects (pieces) located outside the region.
Class for Inner Pieces

Select or create a class. To create a new class, type a name for the new class and click
OK to open
Class for Outer Pieces

Select or create a class. To create a new class, type a name for the new class and click
OK to open.

Reference Book

11 June 2014

9 Advanced Object Reshaping

Algorithms
Advanced reshaping algorithms include sophisticated algorithms supporting a variety of
complex object shape transformations.

9.1
9.1.1

Shape Split (Prototype)

Supported Domains

Execute; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

9.1.2

Algorithm Parameters

Maximum Border Ratio

Measures the relative border length of the cut line for both resulting objects. Cutting is
only executed if the measured value is below this threshold for both objects. Use a zero
value to disable this feature. This parameter is only used for contraction mode.
Maximum Cut Point Distance

Measures the distance of the cutting points on the object border. Cutting is only performed if this value is below this threshold. Enter a zero value to disable this feature.
Maximum Border Length

Restricts the maximum border length of the smaller objects. Enter a zero value to disable
this feature.

Advanced Object Reshaping Algorithms

9.2

Multiresolution Segmentation Region Grow

Grow segmented image objects according to the multiresolution segmentation criteria.

For a detailed description of all parameters, refer to the algorithm Multiresolution Segmentation (p 35).

9.2.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

9.2.2

Image Layer Weights

Enter weighting values the higher the weight assigned to an image layer, the more weight
will be given to that layers pixel information during the segmentation process.
You can also use a variable as a layer weight.

9.2.3

Thematic Layer Usage

Thematic layer usage ags

Scale Parameter

Scale parameter
Candidate Classes

Candidate classes lter

9.2.4

Composition of Homogeneity Criteria

Shape

Weight of the shape criterion

Compactness

Weight of the compactness criterion

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

9.3

Image Object Fusion

Dene a variety of growing and merging methods and specify in detail the conditions for
merger of the current image object with neighboring objects.
Image object fusion uses the term seed for the current image object. All neighboring
image objects of the current image object are potential candidates for a fusion (merging).
The image object that would result by merging the seed with a candidate is called the
target image object.
A class lter enables users to restrict the potential candidates by their classication. For
each candidate, the tting function will be calculated. Depending on the tting mode, 1
one or more candidates will be merged with the seed image object. If no candidate meets
all tting criteria no merge will take place.

Figure 9.1. Example for image object fusion with seed image object S and neighboring objects

A , B , C and D

9.3.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

9.3.2

Candidate Settings

Enable Candidate Classes

Select Yes to activate candidate classes. If the candidate classes are disabled the algorithm
will behave like a region merging.
1. If you do not need a tting function, we recommend that you use the algorithms Merge Region and Grow
Region. They require fewer parameters for conguration and provide higher performance.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

Candidate Classes

Choose the candidate classes you wish to consider. If the candidate classes are distinct
from the classes in the domain (representing the seed classes), the algorithm will behave
like a growing region.

9.3.3

Fitting Function

The fusion settings specify the detailed behavior of the Image Object Fusion algorithm.
Fitting Mode

Choose the tting mode (see table 9.1, Fitting Mode Options for Image Object Fusion
Algorithm).
Table 9.1. Fitting Mode Options for Image Object Fusion Algorithm

Value

Description

All tting

Merges all candidates that match the tting criteria with the seed.

First tting

Merges the rst candidate that matches the tting criteria with the seed.

Best tting

Merges the candidate that matches the tting criteria in the best way with
the seed.

All best tting

Merges all candidates that match the tting criteria in the best way with
the seed.

Best tting if
mutual

Merges the best candidate if it is calculated as the best for both of the two
image objects (seed and candidate) of a combination.

Search mutual
tting

Executes a mutual best tting search starting from the seed. The two
image objects tting best for both will be merged. Note: These image
objects that are nally merged may not be the seed and one of the original
candidate but other image objects with an even better tting

Fitting Function Threshold

Select the feature and the condition you want to optimize. The closer a seed candidate
pair matches the condition, the better the tting.
Use Absolute Fitting Value

Enable to ignore the sign of the tting values. All tting values are treated as positive
numbers independent of their sign.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

9.3.4

Weighted Sum

Dene the tting function. The tting function is computed as the weighted sum of
feature values. The feature selected in Fitting Function Threshold will be calculated for
the seed, candidate, and the target image object. The total tting value will be computed
by the formula. Fitting Value = (Target * Weight) + (Seed * Weight) + (Candidate *
Weight) To disable the feature calculation for any of the three objects, set the according
weight to 0.
Target Value Factor

Set the weight applied to the target in the tting function.

Seed Value Factor

Set the weight applied to the seed in the tting function.

Candidate Value Factor

Set the weight applied to the candidate in the tting function:

Typical Settings (TVF, SVF,
CVF)

Description

1, 0, 0

Optimize condition on the image object resulting from the

merge

0, 1, 0

Optimize condition on the seed image object

0, 0, 1

Optimize condition on the candidate image object

2, 1, 1

Optimize the change of the feature by the merge

9.3.5

Merge Settings

Fusion Super Objects

This parameter denes the behaviour when the seed and candidate objects selected for
merging have different super objects. If enabled, the super objects will be merged with
the sub objects. If disabled the merge will be skipped.
Thematic Layers

Specify the thematic layers that are to be considered in addition for segmentation. Each
thematic layer used for segmentation will lead to additional splitting of image objects
while enabling consistent access to its thematic information. You can segment an image
using more than one thematic layer. The results are image objects representing proper
intersections between the thematic layers.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

Compatibility Mode

Select Yes from the Value eld to enable compatibility with older software versions (version 3.5 and 4.0). This parameter will be removed with future versions.

9.3.6

Classification Settings

Dene a classication to be applied to the merged image objects.

Active Classes

Choose the list of active classes for the classication.

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

If Yes is selected, class descriptions are evaluated for all classes. The image object
is assigned to the class with the highest membership value.
If No is selected, class descriptions are ignored. This option delivers valuable
results only if Active Classes contains exactly one class.
If you do not use the class description, we recommend you use the Assign Class algorithm
instead.

9.4

Border Optimization

Change the image object shape by either adding sub-objects from the outer border to the
image object or removing sub-objects from the inner border of the image object.

9.4.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

9.4.2

Border Optimization Settings

Candidates

Choose the classes you wish to consider for the sub-objects. Sub-objects need to be
classied with one of the selected classes to be considered by the border optimization.
Destination

Choose the classes you wish to consider for the neighboring objects of the current image
object. To be considered by the Dilatation, sub-objects need to be part of an image object
classied with one of the selected classes. To be considered by the Erosion sub-objects
need to be moveable to an image object classied with one of the selected classes. This
parameter has no effect for the Extraction.
Operation

Choose from the following values:

Value

Description

Dilatation

Removes all Candidate sub-objects from its Destination superobject inner border
and merges them to the neighboring image objects of the current image object.

Erosion

Removes all Candidate objects from its Seed superobject inner border and merges
them to the neighboring image objects of Destination domain.

Extraction Splits an image object by removing all sub-objects of the Candidate domain from
the image objects of Seed domain.

9.4.3

Classification Settings

The resulting image objects can be classied.

Active Classes

Choose the list of active classes for the classication.

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

Use Class Description

If Yes is selected, class descriptions are evaluated for all classes. The image object
is assigned to the class with the highest membership value.
If No is selected, class descriptions are ignored. This option delivers valuable
results only if Active Classes contains exactly one class If you do not use the class
description, we recommend you use the Assign Class algorithm instead.

9.5

Morphology

Smooth the border of image objects by the pixel-based binary morphology operations
Opening or Closing. This algorithm refers to image processing techniques based on mathematical morphology.

9.5.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Superobject; Linked
Object; Image Object List

9.5.2

Morphology Settings

Operation

Decide between the two basic operations: Opening or Closing. Conceptually, imagine using Opening for sanding image objects and Closing for coating image objects. Both will
result in a smoothed border of the image object. Open Image Object removes pixels from
an image object. Opening is dened as the area of an image object that can completely
contain the mask. The area of an image object that cannot contain the mask completely
is separated.

Figure 9.2. Opening operation of the morphology algorithm

Close Image Object adds surrounding pixels to an image object. Closing is dened as
the complementary area to the surrounding area of an image object that can completely
contain the mask. The area near an image object that cannot contain completely the mask
is lled; thus comparable to coating. Smaller holes inside the area are lled.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

Mask

Dene the shape and size of mask you want. The mask is the structuring element, on
which the mathematical morphology operation is based. In the Value eld, the chosen
Mask pattern will be represented on one line. To dene the binary mask, click the ellipsis
button. The Edit Mask dialog box opens.

Figure 9.3. Edit Mask dialog box

To modify the binary mask you have the following options:

Change the Width of the mask by entering new positive number.
Create Square helps you to create a quadratic mask. Enter the dimensions. Start
trying with values similar to the size of areas +1 you want to treat by sanding or to
ll by coating. (Square masks perform less-precise operations and produce fewer
artifacts than circle masks do.)
Create Circle helps you to create a circular mask. Enter the side length. Start trying
with values similar to the size of areas +1 you want to treat by sanding or to ll by
coating. Alternatively, you can directly dene a binary mask in the mask text eld
using a full-stop for FALSE and the hash symbol (#) for TRUE.
3.5 Compatibility Mode

Select Yes from the Value eld to enable compatibility with older software versions (version 3.5 and 4.0). This parameter will be removed with future versions.

9.5.3

Classification Settings

When the operation Open Image Object is active, a classication will be applied to all
image objects sanded from the current image object. When using the Close Image Object
operation, the current image object will be classied if it gets modied by the algorithm.
Active Classes

Choose the list of active classes for the classication.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

100

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

9.6

Watershed Transformation

Calculate an inverted distance map based on the inverted distances for each pixel to the
image object border. Afterwards, the minima are ooded by increasing the level (inverted
distance). Where the individual catchment basins touch each other (watersheds), the
image objects are split.
The watershed transformation algorithm is commonly used to separate image objects
from others. Image objects to be split must already be identied and classied.

9.6.1

Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Superobject; Linked
Object; Image Object List

9.6.2

Watershed Settings

Length Factor

The Length factor is the maximal length of a plateau, which is merged into a catchment
basin. Use the toggle arrows in the Value eld to change to maximal length. The Length
Factor must be greater or equal to zero.

9.6.3

Classification Settings

Dene a classication to be applied if an image object is cut by the algorithm.

Reference Book

11 June 2014

Advanced Object Reshaping Algorithms

101

Active Classes

Choose the list of active classes for the classication.

Erase Old Classification If There Is No New Classification

If you select Yes and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is deleted.
If you select No and the membership value of the image object is below the acceptance threshold (see classication settings) for all classes, the current classication
of the image object is kept.
Use Class Description

Reference Book

11 June 2014

10 Pixel-Based Object
Reshaping Algorithms
Pixel-based reshaping algorithms modify the shape of image objects by adding or removing pixel/voxels according to given criteria.

10.1 Pixel-Based Object Resizing

Grow or shrink image objects based on pixel criteria. Typically, the Relative Area of
Classied Objects feature is used to nd suitable seed image objects. These are grown
or shrunken to larger or smaller image objects. In addition you can use this algorithm to
smooth the surface of image objects by growing or shrinking. You can choose to resize
in one, two or three dimensions.

10.1.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

10.1.2 Algorithm Parameters

Resizing modes
Growing

Grow each seed image object. The starting extents of the seed image objects

are lost.

Figure 10.1. Sample starting classification (left) and after 10x growing (right)

102

Pixel-Based Object Reshaping Algorithms

103

Add an new image object around each seed image objects and grows it. The
seed image objects continue to exist as separate image objects with unchanged extent.
Coating

Figure 10.2. Sample starting classification (left) and 10x coating (right). Coating works from

outside the Candidate image objects to the center

Add an new image object inside each candidate image objects and grows it. A
candidate image object shrinks by the extent of the new image object. This mode works
similar like coating, but inside of candidate image objects.
Shrinking

Figure 10.3. Sample starting classification (left) and after 10x shrinking (right). Shrinking
works from outside the Candidate image objects to the center

Class for New Image Objects

Select a class to be assigned to the new image objects. This feature is available for
Coating or Shrinking modes, but not for Growing.
Preserve Current Object Type

Select the type of image object to determine how image objects of the target class are
merged.
Value Description
Yes

Newly created image objects are 2D-connected. Other image objects maintain their
connectivity (default)

Newly created image objects are disconnected. Any overwritten objects or split
sub-objects are disconnected.

Enable compatibility to version < 8.0

Enables the compatibility mode to software versions earlier than 8.0.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

104

10.1.3 Growing/Shrinking Parameters

Dene the directions where growing or shrinking should be allowed.
X direction

Restrict (no) or enable (yes) resizing along the X axis.

Y direction

Restrict (no) or enable (yes) resizing along the Y axis.

Z direction

Restrict (no) or enable (yes) resizing along the Z axis.

10.1.4 Candidate Object Domain Parameters

Dene the set of neighboring image objects whose pixels are considered for growing. A
pixel can be only added to an image object that is part of the candidate object domain.
This feature is available for Coating or Shrinking modes, but not for Growing.
Class Filter

Select a candidate class. Image objects that are selected in the domain are automatically
excluded from the candidates.
Threshold Condition

Dene an additional threshold condition to dene the candidate domain. Only pixels that
belong to image objects that fulll the threshold condition are considered for resizing.

10.1.5 Pixel Level Constraint Parameters

Pixel Layer Constraint 1; Pixel Layer Constraint 2
In addition, you can dene one or two independent image layer intensity conditions to be
fullled by candidate pixels on a given image layer. The candidate pixels are only added
to the set of active pixels if the conditions dened are fullled.
Layer

Select any image layer to be used for the pixel layer constraint.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

105

Operation

Select a comparison operation to be executed. An image layer must be selected.

Value Description
<

Less than. (Default)

Less than or equal to

Equal to

Greater than

Greater than or equal to

Reference

Select the type of value used for comparing to the pixel layer intensity value. An image
layer must be selected
Value

Description

Absolute value Compare with an absolute layer intensity value that you can dene; see Value
below. It also can be represented by a variable.
Value of
current pixel

Compare with the layer intensity value of the current pixel. You can dene a
Tolerance, see below.

Value

Enter a layer intensity value used as threshold for the comparison operation. Alternatively,
you can select a feature or a variable. To create a new variable, type a name for the new
variable and click OK to open the Create Variable dialog box for further settings. An
absolute value has to be selected as a reference option.
Tolerance

Enter a value used as tolerance for the threshold Value of the comparison operation. The
value of the current pixel has to be selected as a reference option. Alternatively, you can
select a feature or a variable.
Tolerance Mode

Select a calculation mode for the Tolerance value of the threshold value of the comparison
operation. The value of the current pixel must be selected as a reference option.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

106

Value

Description

Absolute

The Tolerance value represents a percentage value. For example, 20 means a

tolerance of 20 gray values.

Percentage The Tolerance value represents a percentage value. For example, 20 means a
tolerance of 20% of the threshold value.

10.1.6 Candidate Surface Tension Parameters

Use surface tension options to smooth the border of the resizing object.

Figure 10.4. Sample classification after 10x growing without (left) and with (right) surface

tension

Figure 10.5. Sample classification after 10x coating without (left) and with (right) surface

tension

Figure 10.6. Sample classification after 10x shrinking without (left) and with (right) surface

tension

Additionally, you can edit the settings so that the shapes of image objects are smoothed
with no signicant growing or shrinking.
Surface tension uses the Relative Area of Classied Objects region feature of pixels of a
given class to optimize the image object shape while resizing. Within a cube of given size
(gure 10.8) around the current candidate pixel, the ratio of relative area of seed pixels to
all pixels inside the box is calculated.
If the result is according to a given comparison operation (gure 10.9) with a given value,
the current candidate is classied as seed; otherwise it keeps its current classication.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

107

Figure 10.7. Sample classification after 10x coating without (left) and with 10x smoothing

coating (right) based on surface tension

Figure 10.8. Example of calculation of the surface tension of seed objects (gray) is based on

pixels within a box around the current candidate pixel (blue)

Reference

Choose to use either seed pixels of a current image object or pixels of a given class for
surface tension calculation. Your choice inuences smoothing while resizing.
None

Surface tension is not active.

Surface tension is calculated based on image objects: Within the calculation box,
only pixels of the candidate image objects are mentioned. This allows you to smooth
the border of each seed image object without taking neighbor seed image objects into
account.
Object

Figure 10.9. Object-based surface tension calculation while growing keeps image objects more

separated

Class Surface tension is calculated based on a given class (gure 10.10); within the
calculation box, all pixels of a given candidate class are taken into account. This enables
you to smooth the border around neighboring image objects of the same class.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

108

Figure 10.10. Class-based surface tension calculation while growing smooths multiple image

objects of a class

Class Filter

Select a candidate class used to measure the relative area. A class must be dened as
Reference.
Operation

Select a comparison operation to be executed.

Value Description
>=

Greater than or equal to (default)

Less than

Less than or equal to

Equal to

Greater than

Value

Enter a value for the surface tension calculation. Alternatively, you can select a feature
or a variable. To create a new variable, type a name for the new variable and click OK to
open the Create Variable dialog box for further settings.
Box Size in X and Y

Enter the pixel size of the square box around the current candidate pixel in x - and y
-directions to use for the surface tension calculation. The integer value must be uneven;
even integer values are changed to the next higher integer. You can enter a value as a
threshold. Alternatively, you can select a feature or a variable. To create a new variable,
type a name for the new variable and click OK to open the Create Variable dialog box for
further settings. The default value is 5.
Box Size in Z

Similar to Box Size in X and Y. The default value is 1.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

109

10.1.7 Size Limits Parameters

These parameters limit the sizes of growing and shrinking operations.
Min Object Size

The Minimum Object Size parameter stops an object shrinking when it reaches a minimum size. It is very useful for preventing a shrinking object from disappearing (that is
attaining a size of zero).
Max Object Size

The Maximum Object Size parameter stops an object from growing when it reaches a
maximum size. It is very useful for preventing a growing object from leaking.

10.2 Pixel-Based Density Filter

Create new classied image objects from pixel/voxels of image objects in the domain
according to the class density in the surroundings. You can use this algorithm to smooth
the shape of image objects by removing protrusions.
The class density calculation is based on the Relative Area of Classied Objects Region
feature of pixel/voxels of a given class. For each candidate pixel/voxel of image objects
in the domain, a cubic box of given size is used to calculate the ratio of relative area of
classied pixel/voxels to all pixel/voxels inside the box.

Figure 10.11. Example of calculation of the class (gray) density is based on pixels within a box

around a candidate pixel (blue)

If the class density is according to given comparison operations with a given value the
current candidate pixel/voxel is classied as target class, otherwise it keeps its current
classication. Image objects of the target class are merged with each execution of the
algorithm.

10.2.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

110

10.2.2 Algorithm Parameters

Target Class

Select a class to be assigned to image objects according to density criteria on the following page.
Preserve Current Object Type

Select the type of image object to determine how image objects of the target class are
merged.
Value Description
Yes

Newly created image objects are 2D-connected. Other image objects maintain their
connectivity (default)

Newly created image objects are disconnected. Any overwritten objects or split
sub-objects are disconnected.

10.2.3 Growing and Shrinking Directions Parameters

Growing can be enabled or disabled for each dimension. It can also be partially enabled
so that growing occurs towards positive or negative co-ordinates only.

Figure 10.12. Growing directions for two dimensions

X, Y and Z Directions

Dene the search order around active pixels. You can do this for each co-ordinate axis
separately:

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

111

Value

Description

Yes

Enable resizing.(Default for x and y -directions.)

Only Positive

Restrict resizing in direction of positive co-ordinates only.

Only Negative Restrict resizing in direction of negative co-ordinates only.

Disable resizing. (Default for z -direction.)

Figure 10.13. Example of 10x growing in x -direction is restricted to the positive direction and

in y -direction to the negative direction

Balance Growing Directions by Pixel Size

If you work with a three-dimensional scene that has a different voxel size for the zdirection compared to the x - and y -directions, you can choose to take the voxel resolution into account. This ensures that image objects are resized evenly in all directions.
Resizing must be enabled for z -direction.
Value Description
No

Growing is performed in all directions with the same preference.

Yes

In this case the growing takes the extension of the voxel into account. For each
growing step in the z direction the number of growing steps in the xy direction equals
the ratio of the voxel dimensions, for example the xy extend to z extend.

10.2.4 Density Criteria Parameters

You can dene one to three independent density criteria that need to be fullled by candidate pixel/voxels on a given image layer. Only if all criteria are fullled is the candidate
pixel/voxel classied.
Reference

To activate a density criterion, select whether pixel/voxels of a current image object or

pixel/voxels of a given class are used for class density calculation. That way, you inuence the way of smoothing while resizing:

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

112

Value

Description

None

Density criterion is not active. (Not available for Density Criterion 1)

Object Class density is calculated based on the image object that contains the current
candidate pixel/voxel. Within the calculation box, only pixel/voxels of this image
object, are mentioned. This allows you to smooth the border of the image object
without taking neighbor image objects into account.
Class

Class density is calculated based on a given class (see below). Within the calculation
box, all pixel/voxels of the given class are taken into account. This allows you to
smooth the border around neighboring image objects of the same class.

Class Filter

Select a candidate class to use for calculation of class density. A class must be dened as
Reference.
Operation

Select a comparison operation to be executed:

Value Description
>=

Greater than or equal to. (Default)

Less than

Less than or equal to

Equal to

Greater than

Value

Enter a value, or select a feature, or a variable used a threshold for the class density
calculation. To create a new variable, type a name for the new variable and click OK to
open the Create Region Variable dialog box for further settings.
Box Size in X and Y

Enter the pixel/voxel size of the square box around the current candidate pixel/voxel in x
and y -directions, used for the class density calculation. The integer value must be an odd
number; even-numbered integer values are changed to the next higher integer.
You can enter a value, or select a feature, or a variable used as threshold. To create a
new variable, type a name for the new variable and click OK to open the Create Region
Variable dialog box for further settings. The default value is 1.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

113

Box Size in Z

Similar to Box Size in X and Y. The default value is 1.

10.3 Pixel-Based Shape Processing Filters

Provides a set of different lters for processing image objects with respect to various
shape criteria:
Filter

Description

Pixel Count

Detects and modies low-level pixel patterns.

Corner Pixels Detects the corner pixel of an image object. In conjunction with the Connect lter, it
can be used to cut image objects according to their shape.
Connect

Creates a pixel connection between two linked image objects. Can be used to
connect pixels detected by the Corner Pixels or Hole Cutting Pixels lter and thus to
cut image objects.

Inner Border

Reclassies all pixels at the inner border of an image object.

Outer Border

Reclassies all pixels at the outer border of an image object.

Hole
Cutting
Pixels

Detects pairs of pixels at the inner and outer borders of image objects. In
conjunction with the Connect lter, it can be used to cut holes out of image objects.

10.3.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects

10.3.2 Algorithm Parameters

Target Class

Select a class to be assigned to the processed pixels.

Preserve Current Object Type

Select Yes or No:

Value Description
Yes

Ensures that the image object type of processed image objects does not change due to
pixel removal.

Changes the type of all modied image objects to disconnected, regardless of their
current state. You can use the Convert Image Objects algorithm (p 87) to make sure
processed image objects are assigned the desired image object type after processing.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

114

Filter Type

Select a lter.
The Pixel Count lter detects basic pixel patterns in the 4-neighborhood of
a pixel and reclassies all pixels in the domain where a pattern occurs. It can be useful
to connect border pixels to a 4-connected structure or to detect pixel bridges in image
objects. In addition to the basic parameters mentioned above, you can set the following
parameters:
Pixel Count

The Corner Pixels lter detects the corners of an image object specied
by the image object domain. The detected corners can be concave or convex.
Corner Pixels

The lter enables you to further connect detected concave corner pixels using image
object links. In conjunction with the Connect lter mode, these connections can be used
to create lines along which the image objects can be cut according to their shape. In
addition to the basic parameters mentioned above, you can set the following parameters:
The Connect lter creates a pixel line between two linked image objects. It
checks all image objects specied in the image object domain for image object links to
another image object. If an image object link exists, it creates a one pixel line between
the two image objects according to the lter parameters and deletes the image object link.
Connect

This mode is specically designed to be used in conjunction with the Corner Pixel lter
detection of concave corners.
In addition to the basic parameters mentioned above, you can set the following parameters:
Candidate Class Filter species a class lter for candidate objects. Only pixels that
belong to an object of one of the specied classes will be taken into account for the
connection
Exclude Border Pixels 1 species whether pixels located at the border of areas dened by the Candidate Class Filter parameter are taken into account for the connection:
If the value is Yes, border pixels are not taken into account
If the value is No, all pixels of the specied areas are taken into account
Distance Mode species how distances between two pixels are calculated:
Spatial Distance determines the shortest connection along the candidate pixels that can be found between two image objects. Uses predened values
1 for 4-connected pixels and 2 for diagonally connected pixels to calculate
pixel distance
Color Contrast determines the shortest connection along the candidate pixels
that can be found between two image objects. Uses the difference between the
pixel intensities of two pixels for an image layer to calculate pixel distance.
Image Layer species the layer used to calculate pixel distances. It is only available
if the Distance Mode parameter is set to Color Contrast
1. If you enable the Exclude Border Pixels parameter, it might not be possible to establish a valid connection
between two image objects. To avoid endless loops or other unwanted effects, make sure that the rule set you
use can handle this situation properly.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

115

The Inner Border lter reclassies all neighboring pixels of an image object that touch an inner border of the object. This lter is useful to detect holes within
image objects. There are no additional parameters for this lter.
Inner Border

Figure 10.14. Result of Inner Border filter processing

The Outer Border lter reclassies all neighboring pixels of an image

object that touch the outer border of the object. There are no additional parameters for
this lter.
Outer Border

Figure 10.15. Result of Outer Border filter processing

The Hole Cutting Pixels lter detects inner border pixels with close
proximity to the outer border and related pixels on the outer border.
Hole Cutting Pixels

Figure 10.16. Result of Hole Cutting Pixels filter processing

Both pixels are linked so that you can use the Connect lter to cut out so-detected inner
holes from the object. In order to avoid structures that are quite similar to a nearly closed
hole the lter always detects two pairs of inner and outer pixels that have a maximum
distance to each other.
Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

116

Figure 10.17. Result of subsequent Connect filter processing

Mode Parameters
Corner Type

Value

Specify the type of corner pixels to search for:

Description

Concave Detects and reclassies concave corner pixels.

Convex

Detects and reclassies convex corner pixels.

All

Detects and reclassies both concave and convex corner pixels.

Figure 10.18. Image object (left) and its results after detection of concave (middle) and convex
(right) corner pixels

For each border pixel, the lter analyzes the angle of the border by creating
an angle between the pixel and the adjacent border pixels to the left and to the right. The
leg length describes the length of the angle legs in border pixels. Increase the leg length
value to apply a smoothening to the real image object border. Typical values lie within
the range 2 to 9.
Leg Length

Specify the minimum angle (based on the leg length) that must be
present at a border pixel so that it will be considered as a corner.
Minimum Angle

If you enter a value of 0, every border pixel with an angle different to 180 is considered
as a potential corner. The larger the value, the more acute the minimum angle that is
needed for a corner to be accepted.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

117

Figure 10.19. Angle measurement at a border pixel with leg length 3 pxls (approx. 45 for a

concave corner pixel)

This parameter is only available if Corner Type parameter

is set to Concave. It lets you specify whether detected corners are to be connected using
image object links:
Connect Corners for Cutting

Value Description
Yes

Does not connect corners using image object links.

Detects and connects matching corners using image object links. The connections can
be used as input for further processing with the Connect lter. Unconnected corners are
not instantiated.

This parameter is only available if Corner Type parameter is set to

Concave and if Connect Corners for Cutting parameter is set to Yes.
Max. Cut Length

Specify the maximum length of the direct connection between two corner pixels. Corners
are only be linked for connection if their distance is below this threshold. Values are
measured in pixels. A value of 0 will disable this restriction.
This parameter is only available if Corner Type parameter is
set to Concave and if Connect Corners for Cutting parameter is set to Yes. Specify the
maximum ratio of the direct distance between two corners and the distance along the
object border. Corners are only linked for connection if the ratio is below this threshold.
Cut Length/Border Length

Values must lie within the range 0 to 1. A value of 1 will link all corner pairs, a value of
0 will reject all corner pairs.
The example below shows the result of concave corner pixel detection with a leg length
of 3 and a minimum angle of 35. The Connect Corners for Cutting parameter is set to
Yes; a cut length/border length ratio of up to 0.3 is allowed.
The selected pixel objects (marked in red and green) are connected using an image object
link.
Reference

Specify the reference for pixel pattern detection:

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

118

Figure 10.20. Result of Pixel Corner Pixels filter processing with the settings described above

Value

Description

Object Detects patterns of pixels within the same object.

Class

Detects patterns of pixels assigned to the same class.

This parameter is only available if Reference parameter is set to Class. Specify the class lter used for pixel pattern detection in Class reference mode.
Class Filter

Pixel Pattern

Specify the pattern of target pixels to search for:

Table 10.-14. Pixel Pattern Settings

Value

Description

No neighboring pixels. Reclassies isolated pixels

Exactly one neighboring pixel. Reclassies end points of lines with a

width of one pixel

Exactly two neighboring pixels. Combines the reclassication results of 2

(angle) and 2 (line) modes

2 (angle)

Exactly two neighboring pixels forming an angle with the center pixel.
Reclassies pixels that are part of a diagonal pixel bridge. Use this mode
to convert 8-connected pixel structures into 4-connected pixel structures

2 (line)

Exactly two neighboring pixels forming a line with the center pixel.
Reclassies pixels (except end points) being part of lines with a width of
one pixel

Exactly three neighboring pixels

Exactly four neighboring pixels. Reclassies pixels that are completely

surrounded by target pixels

The example below shows the result of a pixel count lter processing on the blue area
using the 2 (angle) pixel pattern. The orange pixels are used as a reference.

Reference Book

11 June 2014

Pixel-Based Object Reshaping Algorithms

119

Figure 10.21. Result of Pixel Count filter processing with the settings described in table 10.-14

on the preceding page, Pixel Pattern Settings

In the course of pixel processing, the orange pixel structure is converted into a 4connected structure that can be merged into one connected image object.
Class lter for candidate objects. Only pixels that belong to an
object of one of the specied classes will be considered for the shortest path.
Candidate Class Filter

When this option is set to yes, the connection must not go along
pixels at the border of the selected class
Exclude Border Pixels

Search in Z Direction

Search also in Z direction (3D images only)

Select the way pixel distances are computed using spatial distance, color
contrast or color value.
Distance Mode

Reference Book

11 June 2014

11 Linking Operation
Algorithms
Use linking operations algorithms to work with image object links. Image object links can
be used to group multiple image objects in different areas of the image without creating
a common superobject.

11.1 Create Links

This algorithm allows you to create and establish object links between image objects.

Figure 11.1. Overlap calculations of the Create Links algorithm

For z = 1 slices, n and n + 1 are scanned and some overlap area is determined. Usually
parameters x, y and t are zero. It may be useful to have values different from zero if
structures such as road networks or time series are to be analyzed. For x, y,t 6= 0 , the
overlap region gets calculated with respect to the shifted template.

11.1.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Superobject; Subobjects; Linked Objects; Image Object List
120

Linking Operation Algorithms

121

11.1.2 Algorithm Parameters

Link Class

Select a class that will be assigned to the new links. This class represents the group name
of the links and is used to compute statistics of the image objects belonging to this link
class.

11.1.3 Candidate Object Domain Parameters

Class Filter

Specify for the classes of candidate image objects to be linked with the seed image objects.
Threshold Condition

Specify an additional condition for candidate objects. Only image objects meeting the
specied condition will be considered for linking.
Map

You can specify a different map to the map selected in the image object domain. In this
case, image object links between different maps will be created. This can be useful to
express object relations between maps.
Candidate PPO

Select the PPO level to use for the objects from the next parent process. This parameter
is only valid if domain and candidate maps are the same.

11.1.4 Overlap Settings Parameters

These parameters are used to specify the shift in space and time from the seed object to
determine the position of objects to that will be linked.
Link To

Use all to link to all objects that match the conditions

Use best to create the link only to the candidate object that has the largest overlap
in the set of all possible overlapping objects.

Reference Book

11 June 2014

Linking Operation Algorithms

122

Overlap Calculation

Select one of the following options:

Value

Description

Do not use overlap

Overlap calculation is omitted. Link is created to any object matching the

target conditions

Relative to larger
object

The ratio of the overlap area to the area of the larger object (between seed
and target object) is calculated

Relative to smaller
object

The ratio of the overlap area to the area of the smaller object is calculated

Relative to current
object

The ratio of the overlap area to the area of the current objects is calculated

Relative to
candidate object

The ratio of the overlap area to the area of the candidate objects is
calculated

Absolute [in pixels] The absolute overlap area in pixels is calculated

X-Position Shift

Shift in pixels in the x direction ( integer values).

Y-Position Shift

Shift in pixels in the y direction ( integer values).

Z-Position Shift

Shift in pixels in the z direction ( integer values).

T-Position Shift

Shift in pixels in the t direction ( integer values).

Min. Required Overlap

Lower threshold for image object overlap. A link will only be created if the calculated
overlap will exceed the specied threshold. Use 0 to disable this parameter.
If the overlap calculation is set to Do Not Use Overlap, Min. Required Overlap is
not available
If the overlap calculation is set to Absolute [in pixels], enter any integer
If the overlap calculation is set to any other option, enter a percentage represented
by a oat number between 0 and 1.

Reference Book

11 June 2014

Linking Operation Algorithms

123

Figure 11.2. Incoming and outgoing links over multiple time frames. The red circles represent

objects and the green arrows represent links

Transformation Parameter Set

This parameter is only available if the map specied in the image object domain is different to the map specied in the candidate object domain.
If you dont know the transformation parameters, you can use the Image Registration
algorithm to perform an automatic afne registration and store the transformation matrix
in a parameter set. Using this transformation information, you can then link objects that
relate to each other in the original maps.

11.2 Delete Links

Delete all or selected object links. The parameters of this algorithm can be used to lter
out specic links to be deleted. All parameters are cumulative; only links fullling all
specied criteria will be deleted.

11.2.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

11.2.2 Algorithm Parameters

Link Class Filter

Classication lter for links to delete. Only links grouped under the selected classes will
be deleted.
Class Filter

Classication lter for linked image objects. Only links that link to an object of one of
the selected classes will be deleted.

Reference Book

11 June 2014

12 Level Operation Algorithms

Level operation algorithms enable you to add, remove, or rename entire image object
levels within the image object hierarchy.

12.1 Copy Image Object Level

Insert a copy of the selected image objects domain above or below the existing one.

12.1.1 Supported Domains

Image Object Level

12.1.2 Algorithm Parameters

Level Name

Enter the name for the new image object level.

Copy Level

Select whether the copied level is placed above or below the input level specied by the
domain.

12.2 Delete Image Object Level

Delete the image object level selected in the domain.

12.2.1 Supported Domains

Image Object Level

124

Level Operation Algorithms

125

12.3 Rename Image Object Level

Rename an image object level.

12.3.1 Supported Domains

Execute

12.3.2 Algorithm Parameters

Level to Rename

Select the image object level to be renamed.

New Level Name

Select or edit an image object level to be changed, and select or edit the new name for the
level. If the new name is already assigned to an existing level, that level will be deleted.
This algorithm does not change names already existing in the process tree.

Reference Book

11 June 2014

13 Map Operations Algorithms

Map operations algorithms are used for working with multiple maps and can be used for
the following actions:
Local storage of temporary results
Multi-scale image analysis and other workspace automation techniques, which enable you to process image data without using the eCognition Server
Comparison of competing and concurrent analysis strategies on the same image
data in parallel. This enables you to select the best results from each analysis and
combine them into a nal result
Comparison of results on different sets of image data in parallel. This enhances
interactive training approaches and helps to develop robust ruleware.

13.1 Copy Map

Copy a map, or part of a map such as current image object to a new map, or overwrite
an existing map with the copy. 2D, 3D and 4D maps can be copied; the minimum size of
a map to be created is 44 pixels.

13.1.1 Supported Domains

Execute; Current Image Object; Neighbor Image Object; Super Object; Sub Objects;
Linked Objects

13.1.2 Algorithm Parameters

Source Region

Dene a region within the source map. Select or enter the name of an existing region. Alternatively, you can enter the coordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner , and its size [Rx , Ry , Rz , Rt ] . The input pattern is
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] .
Alternatively, you can select a variable. To create a new variable, type a name for the new
variable and click OK or press Enter to open the Create Variable dialog box for further
settings.

126

Map Operations Algorithms

127

Target Map Name

The map to be created by copying. Select a map name from the drop-down list or enter
a new name. If you select an existing map, the copy will overwrite it. Alternatively, you
can create a map variable.
Use Variable As Scale

Specify the scale for copying a map using a variable, rather than dening a numerical
value.
Scale Relative to Main Map

Legacy ag to use scale relative to main map.

Scale

Figure 13.1. Select Scale dialog box

If you choose to resample, the Scale will refer to the original image data. If you choose
the default Use Current Scene Scale, the map copy will have the same scale as the map
(or part of a map) being copied. For example, if the main map is copied to map2 with the
Scale at 50%, and map2 is copied to map3 with the Scale at 50%, map3 will be scaled to
50% of the main map, and not 50% of map2:
1. If you do not want to keep the current scale of the map for the copy, click the
ellipsis button to open the Select Scale dialog box.
2. Selecting a scale different to the current scene scale lets you work on the map copy
at a different magnication/resolution.
3. If you enter an invalid scale factor, it will be changed to the closest valid scale as
displayed in the table below.
4. To change the current scale mode, select from the drop-down list. We recommend
that you use the scaling mode consistently within a rule set as the scaling results

Reference Book

11 June 2014

Map Operations Algorithms

128

may differ. Scaling results can differ depending on the scale mode; for example if
you enter 40, you work at the following scales, which are calculated differently:
Options dialog box setting Scale of the scene copy or subset to be created
Units (m/pixel)

40m per pixel

Magnication

40x

Percent

40% of the resolution of the source scene

Pixels

1 px per 40 px of the source scene

Clockwise Rotation Angle

This feature lets you rotate a map that you have copied, by a xed value or with respect
to a variable.
Resampling

Choose whether to apply Gaussian smoothing:

Value

Description

Smooth Smoothing is applied. If smoothing is applied and downsampling occurs, the

algorithm generates smoothed .tif les next to the project le (.dpr)
Fast

Smoothing is not applied. Select for faster processing.

Image Layers

Select an image layer to use in the new map. If no layer is selected, all will be used.
Layers are copied if downsampling occurs and Smooth is selected in the Resampling
eld.
Camera View to Top Down

If set to Yes then point clouds are converted from the Camera View perspective to the
Top Down perspective in the destination map. Raster layers are not transferred to the
destination map.
Copy Thematic Layers

Select a thematic layer to use in the new map. If no layer is selected, all will be used. Thematic vector layers are always copied and converted to thematic raster layers. Thematic
raster layers are copied if downsampling occurs and Smooth is selected in the Resampling
eld.
Copying thematic vector layers can be performance-intensive because vector layers are
converted to raster layers.
Reference Book

11 June 2014

Map Operations Algorithms

129

Thematic Layers

You may specify a thematic layer to use; if no layer is selected, all layers will be used.
Copy Image Object Hierarchy

Choose whether the image object hierarchy is copied with the map or not:
If Yes is selected, the image object hierarchy of the source map will be copied to
the new map.
If No is selected, only selected image and thematic data will be copied to the new
map.
Preserve Current Object Type

When this option is set to No, then created objects can be unconnected, so one object can
have more than one polygon.
Visiblity Flag

If the value is set to Yes (the default), all maps are available from the map drop-down box.
If it is set to No, the created map can be accessed, but cannot be displayed.
Compatibility Mode

Allows compatibility with previous software versions.

Scale Direction

Choose from one of the following:

Plane only
Plane and z
Plane and time
All directions

13.2 Delete Map

Delete a specied map or maps. There are no algorithm parameters.

13.2.1 Supported Domains

Execute; Maps

Reference Book

11 June 2014

Map Operations Algorithms

130

13.3 Synchronize Map

Copy image objects from one map to another.

13.3.1 Supported Domains

Image Object Level

13.3.2 Algorithm Parameters

Target Map Name

Required. Use the dropdown list to select an existing map or map variable, or create a
new map variable and assign a value to it.
Region

Dene a target region to which image objects are copied. Select or enter the name of an
existing region. Alternatively, you can enter the co-ordinates of a region specied by its
origin (xG , yG ) , which is the lower left corner , and its size [Rx , Ry , Rz , Rt ] . The input
pattern is (xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To
create a new variable, type a name for the new variable and click OK or press Enter to
open the Create Variable dialog box for further settings.
Alternatively, you can create a region by entering numbers with that exact syntax: (origin
x , origin y ) - (extent x , extent y ); for example (10,20) - (100,110).
Level

Required. Select the target image object level in the target map.
Class Filter

Select objects of classes on the target map (these can be overwritten). Click the ellipsis
button to open the Edit Classication Filter dialog box. The default is none, which means
objects of any class can be overwritten.
Threshold Condition

Select a threshold condition. Image objects matching the threshold will be overwritten.
Click the ellipsis button to open the Select Single Feature dialog box. The default is none,
which means objects of any class can be overwritten.

Reference Book

11 June 2014

Map Operations Algorithms

131

Clockwise Rotation Angle

This feature lets you rotate a map that you have synchronized, by a xed value or with
respect to a variable. If you have rotated a copied map using the Copy Map algorithm,
you can restore it with the Synchronized Map algorithm by using the negative value of
the angle of rotation.
Preserve Current Object Type

If Yes is selected, the current object type is preserved for all affected objects
If No is selected, modied objects can become disconnected objects.
Synchronize Complete Hierarchy

When this option is set to yes, then all levels on the target map will be affected.
Compatibility Mode

Allows compatibility with previous software versions.

13.4 3D/4D Settings

Dene the map layout settings for multidimensional data sets.
Because Trimble software handles multidimensional data sets internally as a twodimensional image, you have to dene how these two-dimensional maps represent the
corresponding data set.

13.4.1 Supported Domains

Execute

13.4.2 Algorithm Parameters

Mode

Select a mode:

Reference Book

11 June 2014

Map Operations Algorithms

Mode

132

Description

Noninvasive The map size in x is assumed to be the 2D slice size in x and y. This is used to
compute the number of slices.
Example: Imagine a map with the size 2561024. The noninvasive mode uses the
map size x=256 to determine the slice size y=256. Thus, the map is handled as a
3D map consisting of 1024/256=4 slices of size 256256 each.
2D extent

Enter the size of a single slice or a single frame; see below.

4D layout

Depending on the data set, you can enter the number of slices and the number of
frames; see below.

Slice/Frame Size X & Y

Depending on the dataset, enter the x and y sizes per single slice or frame. (2D Extent
must be selected.)
Number of Slices

Enter the number of slices of the data set. (4D Layout must be selected.)
Number of Frames

Enter the number of frames of the data set. (4D Layout must be selected.)
Distance Between Slices

Enter the distance between slices the distance you enter is relative to the xy resolution.
The default is 0.5. For example:
Slice Distance 1 means x = y = z = 1
Slice Distance 2 means 2x = 2y = 1z
Slice Distance 0.5 means x = y = 0.5z
Time Between Two Frames

Enter the time elapsing between frames. The default is 0.5.

Coordinate of the First Slice

Enter the co-ordinate of the rst slice, which determines the co-ordinates of other slices.
The default is 0.

Reference Book

11 June 2014

Map Operations Algorithms

133

Start Time

Enter the time of the rst frame, which determines the time of other frames. The default
is 0.

13.5 Scene Properties

Select the magnication, pixel size and scene unit for a scene, via a rule set.

13.5.1 Supported Domains

Execute; Maps

13.5.2 Algorithm Parameters

Magnification

If appropriate, enter the magnication settings. This is most often used in microscope
images, where a known magnication was used and the information was not embedded
in the image format. You can also enter a variable in this eld. Magnication can only
be set for a scene that was not rescaled.
Scene Unit

Set the default unit for calculating feature values. Choose from pixels, kilometers,
hectometers, decameters, meters, decimeters, centimeters, millimeters, micrometers,
nanometers, angstroms, inches, feet, yards, statute miles and nautical miles.
Pixel Size

If you wish to link the size of your objects to a known scale (for instance in a geographical
image) enter the scene unit that corresponds to one pixel.

13.6 Rename Map

Rename an already existing map.

13.6.1 Supported Domains

Execute

Reference Book

11 June 2014

Map Operations Algorithms

134

13.6.2 Algorithm Parameters

Map to rename

Select the map to be renamed. The main map cannot be renamed.

New map name

Insert a new map name in quotation marks (New Map Name) or create a new variable.

13.7 Transfer Layer

Allows you to transfer a temporary image, vector or point cloud layer between different
maps.

13.7.1 Supported Domains

Execute

13.7.2 Algorithm Parameters

Input - Layer (source layer)

Select the image or vector layer to be transferred.

Output - Map (destination map)

Select the destination map where the image layer will be copied to.
Output - Name

Insert a name for the new layer in the destination map. If no name is dened the same
name as in the source layer will be used.

Reference Book

11 June 2014

14 Image Layer Operation

Algorithms
Image layer operation algorithms are used to create or to delete image object layers. In
addition, you can use them to apply lters to image layers at the pixel/voxel level.

14.1 Distance Map

The Distance Map algorithm computes the distance of all pixels from the image object
domain objects to the objects from the Distance to parameter class. A layer that contains the distances of pixels to objects of a selected class is created using a ood-lling
algorithm.
In a time series distance is not calculated between frames; only between slices in
the x y direction
In scaled maps the distance corresponds to the map coordinates and is scaled with
the map.

14.1.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects

14.1.2 Algorithm Parameters

Distance to

The Distance to parameter lets you select a class; the distance is then calculated to all
objects belonging to this class. The 8-Neighborhood (p 11) calculation is used as it
allows diagonal distance measurements. If no class is selected, the distance of each pixel
to its next image object border is calculated.

135

Image Layer Operation Algorithms

136

Output Layer

Enter a layer name to be used for output. A temporary layer will be created if there is no
entry in the eld or if the entry does not exist. If an existing layer is selected it will be
deleted and replaced.
Output Layer Visible

Select yes or no to create a visible or hidden image layer.

Compatibility mode

Allows compatibility with previous software versions. In previous versions only the distance to neighboring objects was calculated, now the distance to all image objects belonging to the selected class is considered.

14.2 Create Temporary Image Layer

Create a temporary image layer with values calculated from a selected feature for the
image objects selected in the image object domain.

14.2.1 Supported Domains

Image Object Level

14.2.2 Algorithm Parameters

Layer Name

Select the default name for the temporary image layer or edit it.
Feature

Select a single feature that is used to compute the pixel values lled into the new temporary layer.
Value for Undefined

The value to be set, when feature value is undened.

Reference Book

11 June 2014

Image Layer Operation Algorithms

137

Output Layer

Select yes or no to create a visible or hidden image layer.

14.3 Delete Layer

Delete a selected image layer or thematic layer. One image must be kept in the map. This
algorithm is often used in conjunction with the Create Temporary Image Layer algorithm
to remove this image layer after you nished working with it.

14.3.1 Supported Domains

Execute

14.3.2 Algorithm Parameters

Layer Type

Select a type of layer to be deleted thematic layer or image layer.

Layer

Select a layer to be deleted.

14.4 Convolution Filter

Apply a convolution lter to the image. There are two options; a preset Gaussian smoothing lter (Gauss Blur) and a user-dened kernel. A convolution lter uses a kernel, which
is a square matrix of a value that is applied to the image pixels. Each pixel value is replaced by the average of the square area of the matrix centered on the pixel.

14.4.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

Reference Book

11 June 2014

Image Layer Operation Algorithms

138

14.4.2 Algorithm Parameters

Type

The Gauss Blur is a convolution operator used to remove noise and detail. The Custom
Kernel enables the user to construct a kernel with customized values.
Expression

The formula for Gaussian Blur is:

2
1
x
e 22
G(x) =
2
(Where is the standard deviation of the distribution.)
Advanced Parameter

Displays for Gauss Blur. Enter a value for the reduction factor of the standard deviation.
A higher value results in more blur.
Custom Kernel

Displays only when Custom Kernel is selected. Click the ellipsis button on the right to
open the Kernel dialog box (gure 14.1) and enter the numbers for the kernel.

Figure 14.1. Kernel dialog box

The number of entries should equal the square of the kernel size entered in the 2D kernel
size eld. Use commas, spaces or lines to separate the values.

Reference Book

11 June 2014

Image Layer Operation Algorithms

139

14.4.3 Kernel Parameters

2D Kernel Slice

Enter an odd number only for the lter kernel size. The default value is 3.
Number of Slices

This is available if Type is set to Gauss Blur. Enter the number of slices to be considered
as part of the kernel. If a region is specied in the domain, the algorithm will use the
region values in x slices above and x slices below ( x being the number of slices entered).
If there is no region, the entire area of the slices above and below will be considered
part of the kernel. If there are insufcient slices or regions, only those available will be
considered.
Class Filter

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.4.4 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ), which is the lower left corner and its size [Rx , Ry , Rz , Rt ]. The input pattern is:
(xG , yG , zG ,tG ), [Rx , Ry , Rz , Rt ]. Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Select yes or no to create a visible or hidden image layer.

Reference Book

11 June 2014

Image Layer Operation Algorithms

140

Output Layer Type

Select the data type of the output layer. Available options are:

As input layer
8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned
32-bit signed
32-bit oat

14.5 Layer Normalization

The Layer Normalization algorithm offers two options to normalize images. The linear
normalization lter stretches pixel values to the entire pixel value range. The histogram
normalization changes pixel values based on the accumulated histogram of the image.
The effect is illustrated in the following histograms:

Figure 14.2. Sample histogram of original image

Figure 14.3. Example of historgram after linear normalization

Figure 14.4. Sample of histogram after histogram normalization

Reference Book

11 June 2014

Image Layer Operation Algorithms

141

14.5.1 Supported Domains

Pixel Level

14.5.2 Algorithm Parameters

Type

If Linear is selected, a linear stretch is applied to the layer histogram

If Histogram is selected, a histogram stretch is applied to the layer histogram.

14.5.3 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Output Layer

Select yes or no to create a visible or hidden image layer.

Output Layer Type

Select the data type of the output layer. Available options are:

As input layer
8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned
32-bit signed
32-bit oat

14.6 Median Filter

Use the Median Filter algorithm to replace the pixel value with the median value of
neighboring pixels. The median lter may preserve image detail better than a mean lter.
Both can be used to reduce noise.

Reference Book

11 June 2014

Image Layer Operation Algorithms

142

14.6.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.6.2 Kernel Parameters

2D Kernel Size

Enter a number to set the kernel size in one slice. The default value is 3.
Number of Slices

Enter the number of slices to be considered as part of the kernel. If a region is specied in
the domain, the algorithm will use the region values in x slices above and x slices below
( x being the number of slices entered). If there is no region, the entire area of the slices
above and below will be considered part of the kernel. If there are insufcient slices or
regions, only those available will be considered.
Class Filter

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.6.3 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Enter a name for the output layer or use the drop-down list to select a layer name to be
used for output. If left empty, a temporary layer will be created. If a temporary layer is
selected it will be deleted and replaced.

Reference Book

11 June 2014

Image Layer Operation Algorithms

143

Output Layer Type

Select an output layer type from the drop-down list. Select As Input Layer to assign the
type of the input layer to the output layer.

14.7 Sobel Operation Filter

Creates a layer by applying Sobel operator.

14.7.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.7.2 Kernel Parameters

2D Kernel Size

Enter a number to set the kernel size in one slice. The default value is 3.
Number of Slices

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.7.3 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Reference Book

11 June 2014

Image Layer Operation Algorithms

144

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Select yes or no to create a visible or hidden image layer.

Output Layer Type

Select the data type of the output layer. Available options are:

As input layer
8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned
32-bit signed
32-bit oat

14.8 Pixel Freq. Filter

The Pixel Frequency Filter algorithm scans the input layer and selects the color that is
found in the greatest number of pixels. The frequency is checked in the area dened by
the size of the kernel.

14.8.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Image Layer Operation Algorithms

145

14.8.2 Kernel Parameters

2D Kernel Size

Enter a number to set the kernel size. The default value is 3.

Number of Slices

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.8.3 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Select yes or no to create a visible or hidden image layer.

Reference Book

11 June 2014

Image Layer Operation Algorithms

146

Output Layer Type

Select the data type of the output layer. Available options are:

As input layer
8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned
32-bit signed
32-bit oat

14.9 Pixel Min/Max Filter (Prototype)

Creates layer where each pixel has the difference of kernel max or min value to the center
value.

14.9.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

14.9.2 Algorithm Parameters

Mode

Choose the min/max lter mode diff. brightest to center, diff. center to darkest or diff.
brightest to darkest.

14.9.3 Kernel Parameters

2D Kernel Size

Enter a number to set the kernel size in one slice. The default value is 3.
Number of Slices

Reference Book

11 June 2014

Image Layer Operation Algorithms

147

Class Filter

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.9.4 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] .
Alternatively, you can select a variable. To create a new variable, type a name for the new
variable and click OK or press Enter to open the Create Variable dialog box for further
settings.
Output Layer

Select yes or no to create a visible or hidden image layer.

14.10 Edge Extraction Lee Sigma

The Edge Extraction Lee Sigma algorithm uses a specic edge lter that creates two
individual layers from the original image. One layer represents bright edges, the other
one dark edges.
To extract two layers one with bright and one with dark edges this algorithm must be
applied twice with the appropriate settings changed. If two edge layers are created, it is
important to give them two individual image layer aliases. Otherwise, the rst existing
layer would be overwritten by the second generated layer.

Reference Book

11 June 2014

Image Layer Operation Algorithms

148

14.10.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.10.2 Algorithm Parameters

Sigma

Set the Sigma value. The Sigma value describes how far away a data point is from its
mean, in standard deviations. A higher Sigma value results in a stronger edge detection;
the default value is 5. The sigma value for a given window is:
s
2
= (x) 2
If the number of pixels P within the moving window that satisfy the criteria in the formula
below is sufciently large (where W is the width, a user-dened constant), the average of
these pixels is output. Otherwise, the average of the entire window is produced.

1 w PCenter P 1 +W PCenter
Edge Extraction Mode

If Dark is selected, edges of darker objects are extracted

If Bright is selected, edges of brighter objects are extracted
Input Layer

Use the drop-down list to select the input layer.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] .
Alternatively, you can select a variable. To create a new variable, type a name for the new
variable and click OK or press Enter to open the Create Variable dialog box for further
settings.
Output Layer

Enter a name for the output layer or use the drop-down list to select a layer. If a temporary
layer is selected it will be deleted and replaced.

Reference Book

11 June 2014

Image Layer Operation Algorithms

149

14.11 Edge Extraction Canny

Enhance or extract feature boundaries using Cannys algorithm. Edge extraction lters
may be used to enhance or extract feature boundaries. The resulting layer typically shows
high pixel values where there is a distinctive change of pixel values in the original image
layer.

14.11.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.11.2 Algorithm Parameters

Algorithm

The Canny algorithm is provided.

Lower Threshold

Lower threshold is applied after higher threshold. During the rst step, edges are detected
and pixels with values lower than Higher Threshold are removed from detected edges.
During the nal step, non-edge pixels (those previously removed because values were
less than higher threshold) with values higher than lower threshold are marked as edge
nodes again. After applying the algorithm the rst time, you can check results (edge pixel
values) and the value for the threshold. Usually values for this eld are from 05 the
default is 0.
Higher Threshold

After edges are detected, pixels with values lower than this threshold will not be marked
as edge pixels. This allows removal of low intensity gradient edges from results. After
applying the algorithm once, users can check the results (values of edge pixels) and nd
the correct value for the threshold. Usually values for this eld are from 05 the default
is 0.
Gauss Convolution FWHM

Enter the width of the Gaussian lter in relation to full width at half maximum of the
Gaussian lter. This eld determines the level of details covered by Gaussian lter. A
higher value will produce a wider Gaussian lter and less detail will remain for edge
detection. Therefore, only high intensity gradient edges will be detected by Cannys
algorithm. The range of the eld is 0.000115. The default value is 1.

Reference Book

11 June 2014

Image Layer Operation Algorithms

150

Input Layer

Use the drop-down list to select a layer to use for input.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Use the drop-down list to select a layer to use for output or enter a new name. Output is
32-bit oat. If the name of an existing 32-bit oat temporary layer is entered or selected,
it will be used. If a temporary layer is selected it will be deleted and replaced.
Sample Results

Original Layer

The table below shows results of typical settings:

Lower Threshold: 0

Lower Threshold: 0.3

Higher Threshold: 0
Gauss Convolution
FWHM: 0.2

Higher Threshold: 0.6 Higher Threshold:

Gauss Convolution
0.69
FWHM: 0.2
Gauss Convolution
FWHM: 0.2

14.12 Edge 3D Filter

The Edge 3D Filter algorithm creates a layer with 3D edge information.

Reference Book

11 June 2014

Image Layer Operation Algorithms

151

14.12.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Object; Linked Objects

14.12.2 Algorithm Parameters

Smoothing Scale Factor

This parameter denes the sharpness of the detected edges. The value range is 01
the larger the value, the more sharp an edge is displayed in the resulting image. Smaller
values will progressively blur the detected edges until they are unrecognizable.
Return Option

Edge 3D returns all edge signals

Mean + Edge 3D displays bright edges only (mean plus pixel value)
Mean - Edge 3D displays bright edges only (mean minus pixel value)
Edge Finding Method

Abs. Deviation takes all pixels in the kernel

Abs. Deviation of Pixels Along Color Edges only considers pixels on edges

14.12.3 Kernel Parameters

2D Kernel Size

Enter a number to set the kernel size in one slice. The default value is 3.
Number of Slices

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

Reference Book

11 June 2014

Image Layer Operation Algorithms

152

14.12.4 Layer Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ), which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ), [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Select yes or no to create a visible or hidden image layer.

14.13 Surface Calculation

Use the surface calculation algorithm to derive the slope for each pixel of a digital elevation model (DEM). This can be used to determine whether an area within a landscape is
at or steep and is independent from the absolute height values. There is also an option
to calculate aspect using Horns method.

14.13.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.13.2 Algorithm Parameters

Layer

Select the layer to which the lter will be applied. Gradient Unit and Unit of Pixel
parameters apply to slope calculations only.

Reference Book

11 June 2014

Image Layer Operation Algorithms

153

Algorithm

Slope uses the Zevenbergen-Thorne method

Aspect uses Horns method. 2

Gradient Unit

Available for slope. Select Percent or Degree from the drop-down list for the gradient
unit.
Unit of Pixel Values

Enter the ratio of the pixel height to pixel size.

Input Layer

Use the drop-down list to select a layer for input.

Input Region

Dene a region within the input image layer. Select or enter the name of an existing
region. Alternatively, you can enter the co-ordinates of a region specied by its origin
(xG , yG ) , which is the lower left corner and its size [Rx , Ry , Rz , Rt ] . The input pattern is:
(xG , yG , zG ,tG ) , [Rx , Ry , Rz , Rt ] . Alternatively, you can select a variable. To create a new
variable, type a name for the new variable and click OK or press Enter to open the Create
Variable dialog box for further settings.
Output Layer

Select an output layer type from the drop-down list. Select As Input Layer to assign the
type of the input layer to the output layer.
1. Zevenbergen LW, Thorne CR (1987). Quantitative Analysis of Land Surface Topography. Earth Surface
Processes and Landforms, 12(1):4756
2. Horn BKP (1981). Hill Shading and the Reectance Map. Proceedings of the IEEE, 69(1):1447

Reference Book

11 June 2014

Image Layer Operation Algorithms

154

14.14 Layer Arithmetics

The layer arithmetic algorithm uses a pixel-based operation that enables the merger of
up to four layers by mathematical operations (+, , , ). The layer created displays
the result of this mathematical operation. This operation is performed on the pixel level,
which means that all pixels of the image layers are used. For example, Layer 2 can be
subtracted from Layer 1. This would mean that whenever the same pixel value in both
layers exists, the result would be 0.
Before or after the operation, the layers can be normalized. Furthermore, weights can be
used for each individual layer to inuence the result.

14.14.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.14.2 Algorithm Parameters

Input Layer

Select a layer to be used as the input for the lter.

Minimum Input Value

Enter the lowest value of the value range that will be replaced by the output value. The
default is 0.
Maximum Input Value

Enter the highest value of the value range that will be replaced by the output value. The
default is 255.
Output Value

The value that will be written in the new calculated raster layer. May be a number or
an expression. For example, to add Layer 1 and Layer 2, enter Layer 1 + Layer 2. The
following operations can be used in the expression:

Basic operations (+,-,*,/)

Power operation (^)
Trigonometric operations (sin(), arcsin(), cos(), arccos(), tan(), arctan())
Logarithmic operations (ln(), exp(), lg(), exp10())
Boolean operations (<, =, >)
Logical operations (& |)
Coordinate operations (_x_,_y_)
Floor function (oor())

Reference Book

11 June 2014

Image Layer Operation Algorithms

155

Constant (pi)
Formula Examples for Output Value

Layer 1+(Layer 2*10)

(Layer 1>128)*255
(Layer 1<0.5)*Layer 2+(Layer 1=0.5)*Layer 3+(Layer 1>0.5)*Layer 3
Layer 1^0.5-Layer 2+Layer 3/8
sin(Layer 1)
"_x_"

For more examples have a look here: http://www.ecognition.com/cynin_login?ajax

=true&back=http://community.ecognition.com/home/Layer%20Arithmetics%20sample
%20Rule%20Set.zip/view This example project shows you how to apply the algorithm
layer arithmetrics.
Note the spacing for operators and layer names. Use an expression (in quotation marks)
to avoid creating variables, such as a variable with the name Layer 1. If a local layer needs
to be referenced, it is necessary to manually enter the full expression (e.g. CA.Layer 1)
in the eld.
If variables have previously been created (for example, a user accidentally enters Layer
1 without quotation marks), the value of the variable will be used even if the same name
is set in quotation marks subsequently. You will need to delete the variable or edit it in
order to use the layer in the Output Value eld.
Output Layer

Region within the output layer.

Output Layer Visible

Select yes or no to create a visible or hidden image layer.

Output Layer Type

This eld is available if the output layer does not yet exist. Select a data type for the raster
channel if it must be created:

8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned

Reference Book

11 June 2014

Image Layer Operation Algorithms

156

32-bit signed
32-bit oat

14.15 Line Extraction

The line extraction algorithm creates a layer and classies the pixels of the input layer
according to their line lter signal strength.

14.15.1 Supported Domains

Pixel Level

14.15.2 Algorithm Parameters

Line Direction

Enter the direction of the extracted line in degrees, between 0 and 179. The default value
is 0.
Line Length

Enter the length of the extracted line. The default value is 12.
Line Width

Enter the length of the extracted line. The default value is 4.

Border Width

Enter the width of the homogeneous border at the side of the extracted line. The default
value is 4.
Max Similarity of Line to Border

Enter a value to specify the similarity of lines to borders. The default value is 0.9.
Min Pixel Variance

Enter a value to specify the similarity of lines to borders. Use 1 to use the variance of
the input layer. The default value is 0.

Reference Book

11 June 2014

Image Layer Operation Algorithms

157

Min Mean Difference

Enter a value for the minimum mean difference of the line pixels to the border pixels. If
positive, bright lines are detected. Use 0 to detect bright and dark lines.
Input Layer

Use the drop-down list to select the layer where lines are to be extracted.
Output Layer

Enter or select a layer where the maximal line signal strength will be written. If a temporary layer is selected it will be deleted and replaced.
Output Layer Visible

Select yes or no to create a visible or hidden image layer.

14.16 Abs. Mean Deviation Filter (Prototype)

The Abs. Mean Deviation Filter algorithm creates an image layer, where each pixel has
the absolute deviation from mean with Gaussian weight.

14.16.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

14.16.2 Algorithm Parameters

Gauss Reduction

Reduction factor of standard deviation along the (unrotated) x axis

Return Option

Select the lter response option:

Circular
Elliptic:
Gauss reduction ( y axis) is the reduction factor of the standard deviation
along the (unrotated) y axis
Rotation angle ( xy axis) is the y axis rotation angle in the xy plane

Reference Book

11 June 2014

Image Layer Operation Algorithms

158

Elliptic difference:
Gauss reduction ( y axis) is the reduction factor of the standard deviation
along the (unrotated) y axis
Rotation angle ( xy axis) is the y axis rotation angle in the xy plane

14.16.3 Kernel Parameters

2D Kernel Slice

Enter an odd number only for the lter kernel size. The default value is 3.
Number of Slices

Species the class lter for source pixels. Pixels outside selected class objects are assumed to have 0 value.

14.16.4 Layers Parameters

Input Layer

Select a layer to be used as the input for the lter.

Input Region

Reference Book

11 June 2014

Image Layer Operation Algorithms

159

Output Layer Visible

Select yes or no to create a visible or hidden image layer.

Output Layer Type

Select the data type of the output layer. Available options are:

As input layer
8-bit unsigned
16-bit unsigned
16-bit signed
32-bit unsigned
32-bit signed
32-bit oat

14.17 Contrast Filter (Prototype)

14.17.1 Supported Domains
Pixel Level

14.17.2 Algorithm Parameters

Input Layer

Input Layer
Temp Channel Alias

Name of generated temporary channel.

Calc Mode

Calculation Mode.
Channels Are Temp

New channels may be temporary or added to the project.

14.18 Pixel Filters Sliding Window

Creates a ltered image layer using various sliding window lter methods along orthogonal axes x, y or z .
Reference Book

11 June 2014

Image Layer Operation Algorithms

160

14.18.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Objects; Sub Objects; Linked Objects

14.18.2 Algorithm Parameters

Algorithm

Choose the algorithm used in ltering.

Input Image Layer

Select the image layer used for ltering. For some object-based lters, no input layer is
needed.
Filtered Layer

Float-valued image layer which stores the lter results.

14.18.3 Filter Kernel Sizes Selection Parameters

Kernel Size in x

Kernel size used for sliding window lter, which should always be an odd number. If the
kernel size is 1, this lter axis is not used.
Kernel Size in y

Kernel size used for sliding window lter, which should always be an odd number. If the
kernel size is 1, this lter axis is not used.
Kernel Size in z

Kernel size used for sliding window lter, which should always be an odd number. If the
kernel size is 1, this lter axis is not used.

14.19 Fill Pixel Values

The Fill Pixel Values algorithm lets you ll in raster regions based on the domain, by
interpolation from neighbor object pixels. It is generally suitable for interpolating nondata areas in continuously varying rasters (such as elevation models).
To use this function, it is necessary to create and classify objects that dene the pixels to
be interpolated (such as those with a zero value). This function works only on temporary

Reference Book

11 June 2014

Image Layer Operation Algorithms

161

image layers these can be created or copied using the Layer Arithmetics on page 154
algorithm).

14.19.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Objects; Sub
Objects; Linked Objects

14.19.2 Algorithm Parameters

Input Layer

Select the temporary image layer to be modied.

Class Filter

Object classes with known pixel values.

Calculation Mode

The following modes are available:

Inverse Distance Weighting
Bi-linear Interpolation (a legacy function)
IDW Distance Weight A general form of nding an interpolated value u at a given point
x based on samples ui = u(xi ) for i = 0, 1, ..., N using IDW is an interpolating function:

1
(uui )2 p
(x, u) = Ni=0 d(x,x
p
i)

derived from the minimizing condition

(x,u)
u

x denotes an interpolated (arbitrary) point, xi is an interpolating (known) point, d is a

given distance metric operator) from the known point xi to the unknown point x . N is the
total number of known points used in interpolation and p is a positive real number, called
the power parameter.
IDW Distance Weight

This value corresponds to the exponent p in the formula for IDW distance weight.

Reference Book

11 June 2014

Image Layer Operation Algorithms

162

14.20 Template Matching

The Template Matching algorithm applies templates created with eCognitions Template
Editor dialog (View > Windows > Template Editor) to image layers. For more details
please refer to the description of the Template Matching (p 66) Editor and see also the
template matching demo projects provided in the community http://www.ecognition.com
/community.

14.20.1 Supported Domains

Pixel Level

14.20.2 Algorithm Parameters

Template Folder

Select the folder path to an existing template.

Input Layer

Select the image layer to be tested against the template.

Output Layer

Name of the raster layer that will be created by the template matching algorithm. The
values of this layer will lie between 1 and 1 and correspond to the correlation coefcient
of the image and the (centered) template at that position.
Threshold

Insert a template match threshold in a range from 1 to 1 with 1 meaning perfect match.
A good indication for the value to use can be obtained in the Test Template tab of the
template editor.
Thematic layer

Name of the thematic layer with template matches marked as points. This thematic layer
has two thematic attributes:
TemplateID: corresponds to 0, 1, etc. depending which template of a group
achieved the best match.
TemplateName: name of the .tif le of the template that achieved the best match.

Reference Book

11 June 2014

15 Point Cloud Algorithms

15.1 LiDAR File Converter
The LiDAR File Converter algorithm lets you extract additional information from loaded
LiDAR data (.las les), such as elevation and number of returns. This information will
be added to the project as additional raster layers.

15.1.1 Supported Domains

None

15.1.2 Algorithm Parameters

Input Data
Image Layer (LAS File Source)

Dene the point cloud layer from which to extract infor-

mation.
Converter Parameters

Point property parameters dene the information to be used when generating new (raster) image layers. The following parameters can be set:
Point Property

Intensity: The point intensity values

X coordinate: X coordinate value of each point
Y coordinate: Y coordinate value of each point
Elevation: The point elevation values
Class: The class information stored in the points
Return number: The Return Number is the pulse return number for a given output
pulse
Number of returns: The Number of Returns is the total number of returns for a
given pulse
Red: The Red image channel value associated with this point
Green: The Green image channel value associated with this point
Blue: The Blue image channel value associated with this point
Number of points: Number of points that are mapped to the pixel

163

Point Cloud Algorithms

164

Distance to camera: Calculates the distance values to the camera position. The
distance unit is according to the unit of the point cloud projection.
Since the data is resampled into a 2D grid with a xed spacing, several LiDAR points can be within one pixel. The calculation mode allows allows computation of
the resulting pixel value out of several LiDAR points. The following modes are available:
Result Mode

Average: The (mean) average of all point properties is calculated

Standard Deviation: standard deviation of all point properties is calculated
Minimum: The minimum value of all available point properties is used
Maximum: The maximum value of all available point properties is used
Median : The median of all available point properties is used
Mode: The mode of all available point properties is used
Closest to camera (camera view only): Only the closest points to the camera are
used for calculation

Because data is resampled into a 2D grid with xed spacing, several points
can be within one pixel. The Returns parameter lets you dene the points to use when
generating the raster image. The following settings are available.
Returns

First: Only the rst return points are used

Last: Only the last return points are used
All: All points are used
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Point Cloud Algorithms

Ignore points with the specified value of blue value

165
Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature. . .
From Array Item. . .
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature. . .
From Array Item. . .
Output Data
Output Layer

The output layer denes the name of the new raster layer.

15.2 Merge point clouds

Merges two point cloud layers into one new, temporary point cloud layer. The new created
point cloud layer can then be used for point cloud segmentation, classication and even
for export.

15.2.1 Supported Domains

execute

15.2.2 Algorithm Parameters

Image Layer

Dene the two point cloud layers you want to merge and set the naming of the output
layer

Reference Book

11 June 2014

Point Cloud Algorithms

166

Image Layer (Las file source)

Select the rst point cloud layer

Image Layer (Las file source)

Select the second point cloud layer

Image Layer Alias

Insert the name of the new created point cloud layer, keep the default name Merge or
enter an individual name. Make sure to set double-quotes. If you select the same name
as one of the input layers it replaces the original input layer.

15.3 Create temporary point cloud

The algorithm lets you create a modied point cloud layer from loaded point cloud data
according to your settings, this layer will be added to the project only temporarily.

15.3.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

15.3.2 Algorithm Parameters

Input
Layer (source point cloud)

Select the point cloud layer from which to create the new

point cloud information.

Point Filter - By Returns

Select which returns should be used for the calculation. The following properties are
available.
Returns

All: All points are used

First: Only the rst return points are used
Last: Only the last return points are used
Single: Uses only the points which belong to the pulse with one return
All of Many: Uses all points from the pulse with multiple returns
First of Many: Only the rst points from pulse with multiple returns
Last of Many: Only the last points from the pulse with multiple returns

Reference Book

11 June 2014

Point Cloud Algorithms

167

Point Filter - By Class

Class

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By Image Layer Comparison
Point Attribute

Dene the point attribute used for comparison.

<none>
Intensity: The point intensity values
Elevation: The point elevation values.
Select the image layer that will provide values to be compared with point
attributes in the .las le.
Image Layer

Upper Value

Upper value of the range.

Lower Value

Lower value of the range.

Point Filter - By Elevation

You can dene to ignore elevation below the specied value.

Enter the value manually or choose:
Minimum allowed elevation

<none>
From Feature
From Array
You can dene to ignore elevation above the specied value.
Enter the value manually or choose:
Maximum allowed elevation

<none>
From Feature
From Array
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Point Cloud Algorithms

Ignore points with the specified green value

168
Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Output

Dene the name of the new temporary point cloud layer. The default
name is temporary. If you select one of the image layers from the drop down menu the
respective layer will be overwritten.
Output Layer

Select a map for the new temporary point cloud layer. The default setting <Use map from Domain> inserts the temporary point cloud to the Domain specied.
Map (output map)

15.4 Transfer temporary point cloud information

Transfers point information from a temporary point cloud layer to a permanent destination
point cloud layer.

15.4.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

15.4.2 Algorithm Parameters

Input - Layer (temporary point cloud)

Select the temporary point cloud source layer.

Input - Layer (destination point cloud)

Select the name of the permanent destination point cloud layer.

Reference Book

11 June 2014

Point Cloud Algorithms

169

15.5 Export point cloud

This algorithm lets you export point cloud (.las) les with point re-classication.

15.5.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

15.5.2 Algorithm Parameters

Input
Image Layer (point cloud source file)

Dene the point cloud layer from which to export

the .las le.

Point Filter - By Returns
Returns:

The returns parameter lets you dene the points to use when exporting the .las

le.

All: All points are used

Point Filter - By Class

Class

The following options are available:

<none>: No ltering applied this is the default setting

Selected Classes: All operations are applied to points of the selected classes
All Not Selected Classes: All operations are applied to points of the classes that
are not selected.
If ltering is applied, a list of classes standardized for the .las le format is provided.
Standard classes allow the use of predened standardized class types; for custom classes,
other selection classes are available.
Point Filter - By Image Layer Comparison
Point Attribute

Dene the point attribute used for comparison.

<none>
Intensity: The point intensity values

Reference Book

11 June 2014

Point Cloud Algorithms

170

Elevation: The point elevation values.

Select the image layer that will provide values to be compared with point
attributes in the .las le.
Image Layer

Upper Value

Upper value of the range.

Lower Value

Lower value of the range.

Point Filter - By Elevation

You can dene to ignore elevation below the specied value.

Enter the value manually or choose:
Minimum allowed elevation

<none>
From Feature. . .
From Array Item. . .
You can dene to ignore elevation above the specied value.
Enter the value manually or choose:
Maximum allowed elevation

<none>
From Feature. . .
From Array Item. . .
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature. . .
From Array Item. . .
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature. . .
From Array Item. . .
Output - LiDAR Class

Dene the class with which the points are labeled with or use the option leave original
class(es) to store original class(es).

Reference Book

11 June 2014

Point Cloud Algorithms

171

Output - Export Type

Choose whether to overwrite the original input le, or create a new one. The options are:
classify points in original LAS le
classify points in new LAS le (complete point cloud)
classify points in new LAS le (subset only)

15.6 Export line from point cloud

Users classied points from the selected point cloud to create a polyline for each object
of image object information.

15.6.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects

15.6.2 Algorithm Parameters

Feature Parameters
Image Layer (LAS file source)

Dene the point cloud layer from which to export the .las

le.
Detail(0. . . 1)

Dene with a value from 0 to 1 how detailed the export area is

Smoothness

Dene how smooth the line is exported Enter a value or choose

number of pulses
performance with precalculation
performance without precalculation
choose value from fature
choose value from array

Point Filter
Returns

The returns parameter lets you dene the points to use when exporting the .las

le.
All: All points are used
First: Only the rst return points are used
Last: Only the last return points are used

Reference Book

11 June 2014

Point Cloud Algorithms

172

Classification Filter
Filtering

The following options are available:

<None>: No ltering applied this is the default setting

Static Export Item writes the export item to the workspace

Enter a name in Export Item Name, or use the default name
Dynamic Export Item lets you enter or select a variable as an export item
Select a variable from the drop-down box or enter a name in the Export Item
Variable Name eld (entering a name launches the Create Variable dialog box,
where you can enter a value and variable type)
Use Explicit Path exports to the location dened in the Export Path eld. The
default export path is {:Workspc. OutputRoot} \results\{:Item. Name}
\{:Project. Name}.v {:Project.Ver}.{:Ext}

15.7 Cluster analysis of point cloud

This algorithm tries to nd areas of similar pattern, i.e. similar elevation. It classies
directly in the original point clouds.

15.7.1 Supported Domains

Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super
Object; Sub Objects; Linked Objects; Image Object List

15.7.2 Algorithm Parameters

Feature Parameters
Image Layer(Las file source)

Select the las le you want to analyze

Point Filter
Use property

Choose from which property the pattern shall be calculated

Intensity: The point intensity values

Reference Book

11 June 2014

Point Cloud Algorithms

173

Elevation: The point elevation values

Class: The class information stored in the points
Return number: The Return Number is the pulse return number for a given output
pulse
Red: The Red image channel value associated with this point
Green: The Green image channel value associated with this point
Blue: The Blue image channel value associated with this point
Number of clusters

Neighbors

Choose in how many clusters the points shall be divided

Choose the search radius to calculate the pattern

Minimum allowed elevation

Set a minimum value

Maximum allowed elevation

Set a maximum value

LiDAR Class

Dene the class with which the points in one cluster are labeled with

Input set filter

Filtering

Target set filter

Filtering

Reference Book

11 June 2014

16 Thematic Layer Operation

Algorithms
Thematic layer operation algorithms are used to transfer data from thematic layers to
image objects and vice-versa.

16.1 Create Temporary Thematic Vector Layer

Creates a temporary thematic vector layer.

16.1.1 Supported Domains

Image Object Level

16.1.2 Algorithm Parameters

Output layer name

Select the name of the new temporary vector layer. The default name is temporary.
Output Map

Select a map for the new temporary thematic layer. The default setting <Use map from
Domain> inserts the temporary layer to the Domain specied.
Output vector type

Type of the vector layer to create (only 3D point is supported in this version).
Output layer visible

Select to create visible (Yes) or hidden (No) image layer.

174

Thematic Layer Operation Algorithms

175

Vector content

Type of the source data for vector layer contents. The following selections are possible:
Create empty
Create from point cloud
When selecting Create from point cloud the following options are available:
Source point cloud data
Source point cloud

Species which input point cloud le should be used for calculation.

Select the statistical operation to compute the X coordinate from the point
cloud with the statistical operation applied per image object.
X coordinate

Select the statistical operation to compute the Y coordinate from the point
cloud with the statistical operation applied per image object.
Y coordinate

Select the statistical operation to compute the Z coordinate from the point
cloud with the statistical operation applied per image object.
Z coordinate

Point Filter - By Class

Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

Ignore points with the specified value of blue value

176
Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature. . .
From Array Item. . .
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature. . .
From Array Item. . .
Rasterization Settings - Point size

Point size in pixels when rasterizing vector point.

Attribute Table - Attribute table

Select the attributes to be attached.

16.2 Assign Class by Thematic Layer

Assign the classes stored in a thematic layer to image objects. This algorithm quickly
assigns all classes in thematic layers to an image object level. If an image object contains pixels with different classication within the thematic layer, the classication of the
majority of the pixels is assigned to the image object.
Classes are matched by name between the thematic layer and the class hierarchy. The
algorithm assumes that classes with the same names must be present in the class hierarchy.
The class mode parameter sets the behavior if no matching class can be found.

16.2.1 Supported Domains

Image Object Level; Image Object List

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

177

16.2.2 Algorithm Parameters

Thematic Layer

Specify the thematic layer where classication values will be read.

Thematic Layer Attribute

Specify the thematic layer attribute column that contains classication values.
Class Mode

Specify the behavior when the class specied in the thematic layer is not present in the
rule set.
Skip if class does not exist class assignment is skipped if the class specied in
the thematic layer is not present in the rule set
Use default class use a default class if the class specied in the thematic layer is
not present in the rule set
Create new class a random, unique color will be assigned to the new class in the
Class Hierarchy window.
Default Class

Only available if class mode is Use Default Class. Specify the default class for this mode.

16.3 Synchronize Image Object Hierarchy

Change an image object level to exactly represent the thematic layer. Image objects
smaller than the overlapping thematic object will be merged; image objects intersecting
with several thematic objects will be cut.

16.3.1 Supported Domains

Image Object Level; Image Object List

16.3.2 Algorithm Parameters

Thematic Layer

Select a thematic layer to be represented.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

178

16.4 Read Thematic Attribute

Create and assign local image object variables according to a thematic layer attribute
table. A variable with the same name as the thematic attribute will be created, attached
to each image object in the domain and lled with the value given by the attribute table.

16.4.1 Supported Domains

Image Object Level; Image Object List

16.4.2 Algorithm Parameters

Thematic Layer

Select the thematic layer for the algorithm.

Thematic Layer Attribute

Choose attributes from the thematic layer for the algorithm. You can select any numeric
attribute from the attribute table of the selected thematic layer.
Variable

Select an existing variable or enter a new name to add a new one. If you have not already
created a variable, the Create Variable dialog box will open.

16.5 Write Thematic Attributes

Generate a attribute column entry from an image object feature. The updated attribute
table can be saved to a .shp le.

16.5.1 Supported Domains

Image Object Level

16.5.2 Algorithm Parameters

Thematic Layer

Select the thematic layers for the algorithm.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

179

Feature

Select the feature for the algorithm.

Save Changes to File

If the thematic layer is linked with a shapele, the changes can be updated to the le.

16.6 Modify Thematic Attribute Column

This algorithm updates thematic layer attribute columns. Attribute columns can be added,
removed or renamed.

16.6.1 Supported Domains

Execute

16.6.2 Algorithm Parameters

Thematic Layer

Select the thematic layer for the algorithm.

Mode

Select the algorithm mode for the algorithm:

Add Column

Column name: Enter column name to add. Consider that shape les (*.shp) do not
support more than 9 characters.
Column type: Choose column type (double, integer, string)
Default value: Specify default value
Rename Column

Existing column: Enter existing column name

New name: Choose new column name
Remove Column

Existing column: Enter existing column name to be deleted.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

180

16.7 Convert Image Objects to Vector Objects

This algorithm converts raster image objects to polygon, line or point vector objects. In a
second step these vectors can be saved using an export algorithm or added to an existing
vector layer (e.g. Vector Boolean Operation Algorithm).

16.7.1 Supported Domains

Image Object Level

16.7.2 Algorithm Parameters

Thematic layer

Select a name for the thematic layer to be created. The default name is converted_objects.
Shape type

Select the shape type to be created:

Polygons
Lines
Points
Attributes

Select the features for the vector attribute table.

16.8 Vector Boolean Operation

This algorithm performs an overlay analysis by applying boolean operations based on the
vector layers from the domain. There are four boolean operations (union, intersection,
substract and difference) available. The results are appended to the output layer. The
algorithm deletes content from the output layer in case of a different geometry type.

16.8.1 Supported Domains

Vectors (multiple layers)

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

181

16.8.2 Algorithm Parameters

Boolean operation

Select the boolean operation to be applied:

Union
Intersection
Difference
Substraction

Substract vector layer

Only available for the boolean operation substraction to select the vector layers to be
substracted from the domain vector layers.
Output vector layer

Select the name of the output vector layer. Dependent on the selected operation the default names are vector_union, vector_intersection, vector_difference or vector_substract.
Overwrite output layer

The default setting to overwrite the output layer is No. If the output vector layer exists
already and you want to overwrite it select Yes.

16.9 Vector Smoothing

This algorithm generalizes line or polygon vector layers. The two different algorithms
available are Polynomial curve and Cubic Bzier curve. (See also algorithm vector simplication.)
The Polynomial smoothing algorithm (Hermite Interpolation) uses third-degree polynomials with tangents calculated using cardinal splines. The interpolated lines always pass
through the polygon vertices.
The Cubic Bezier smoothing applies Bezier curves where rst and last points are end
points of the curve and the start of the curve is tangent to the rst segment of the Bzier
polygon. The control points are adjusted so that the curve goes through the line or polygon vertices.

16.9.1 Supported Domains

Vectors

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

182

16.9.2 Algorithm Parameters

Output Vector Layer

Insert the output vector layer name for the smoothed vectors. The default name is
vector_smoothed.
Approximation tolerance

The Approximation tolerance is a discretization parameter, i.e. the distance between

points in the approximating curve. Both smoothing algorithms produce polylines. This
approximation parameter inuences the granulation of this polyline and therefore its curvature. The smaller the value the smoother the resulting line will be - on the other hand
more points will be produced which can affect the performance also for further use. An
appropriate value should be small enough so that there is a visual smoothing effect. The
default value in scene units is 0.1.

16.9.3 Smooth mode

Select the respective smoothing algorithm:
Polynomial curve
Cubic Bezier curve

16.10 Vector Simplification

This algorithm simplies lines or polygons using the Douglas-Peucker algorithm. The algorithm reduces the number of points in a curve approximated by a set of points. The goal
of the algorithm is nding a similar curve with fewer points. The degree of simplication
is controlled by the maximum distance between the original shape and the approximated
vectors or based on the percentage of points kept from the original shape.

16.10.1 Supported Domain

Vectors

16.10.2 Algorithm Parameters

Output vector layer

Select the output vector layer name for the simplied vectors. The default name is
vector_simplied.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

183

Stop criteria

Select the Douglas-Peucker algorithm stop criteria based on:

Max distance between node and segment or
Percentage - simplify until specied percentage of original shape.
Epsilon

Insert the Douglas-Peucker distance parameter in scene units. (Only available for Stop
criteria - Max distance.)
Percentage

Percentage of points which should be kept from the original line or polygon shape. (Only
available for Stop criteria - Percentage.)

16.11 Vector Orthogonalization

This algorithm generalizes polygons into rectilinear (orthogonal) polygons based on a
specied granularity.

16.11.1 Supported Domain

Vectors

16.11.2 Algorithm Parameters

Output vector layer

Select the output vector layer name for the generalized polygons. The default name is
vector_rectilinear.
Chessboard size (pixels)

Select the chessboard size in pixels. The algorithm applies a chessboard segmentation
within a bounding box of each polygon assigning each tile either to the polygon or to the
background depending on the merge ratio.
Merge threshold

This threshold denes if a chessboard tile belongs to the polygon or not, it is applied to
the ratio of the number of polygon pixels and number of outside pixels. The value is in a
range between 0 and 1.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

184

Main directon

Polygon direction vector angle with X axis in degrees (range 90,90). The resulting
polygon will be orthogonal and rotated at this angle. If auto-detect option is selected,
then direction is determined by the largest sum of edges with the same angle.

16.12 Vector Integration

This algorithm integrates polygons into a specied vector layer within a specied snapping distance.

16.12.1 Supported Domain

Vectors (multiple layers)

16.12.2 Algorithm Parameters

Algorithm mode

Select the vector integration mode:

Averaging: calculates the average of all points within the snapping distance and
replaces them:
by an average point if no xed vector layer is selected or
snaps to the closest point of a selected xed vector layer
Snapping: snaps every polygon point to the closest point of a xed vector layer
within specied distance, without further action. Precondition for this option is the
selection of a xed vector layer.
Fixed vector layers

Vectors from specied layers will not be changed.

Output vector layer

Select the output vector layer name. The default name is vector_integrated.
Overwrite output layer

Flag to overwrite the output vector layer in case it is already existing.

Snap distance

Snap distance in geo units. This is the largest allowed distance between 2 vectors.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

185

16.13 Vector Remove Intersections

This algorithm removes intersections for polygons. The intersecting area is assigned to
the poylgon with the longer common border.

16.13.1 Supported Domain

Vectors

16.13.2 Algorithm Parameters

Output vector layer

Select the output vector layer name. The default name is vector_clean.

16.14 Vector Dissolve

This algorithm merges vectors based on overlap or common border.

16.14.1 Supported Domain

Vectors (multiple layers)

16.14.2 Algorithm Parameters

Output vector layer

Select the output vector layer name. The default name is vector_dissolve.
Overwrite output layer

Flag to overwrite the output vector layer in case it is already existing.

Criteria Attribute

Select the attribute to use as a merge criteria. An empty value eld results in dissolving
all polygons.

16.15 Delete Vector Objects

This algorithm deletes vector objects of a temporary vector layer.

Reference Book

11 June 2014

Thematic Layer Operation Algorithms

186

16.15.1 Supported Domain

Vectors

Reference Book

11 June 2014

17 Workspace Automation
Algorithms
Workspace automation algorithms are used for working with subroutines of rule sets.
These algorithms enable you to automate and accelerate the processing of workspaces
with particularly large images. Using workspace automation algorithms you can create
multi-scale workows, which integrate analysis of images at different scales, magnications, or resolutions.

17.1 Create Scene Copy

Create a scene copy that is a duplicate of a project with image layers and thematic layers, but without any results such as image objects, classes, or variables. This algorithm
enables you to use subroutines.

17.1.1 Supported Domains

Execute; Maps

17.1.2 Algorithm Parameters

Scene Name

Edit the name of the scene copy to be created.

Scale

See Scale (p 127) for an explanation of the Select Scale dialog box.
Additional Thematic Layers

Edit the thematic layers you wish to load to a scene copy. This option is used to load
intermediate result information that has been generated within a previous subroutine and
exported to a geocoded thematic layer. Use semicolons to separate multiple thematic
layers, for example, ThematicLayer1.tif;ThematicLayer2.tif.
187

Workspace Automation Algorithms

188

17.2 Create Scene Subset

Copy a portion (subset) of a scene as a project with a subset of image layers and thematic
layers. Image objects, classes and variables are not copied. The algorithm uses the given
co-ordinates (geocoding or pixel co-ordinates) of the source scene. You can create subset
copies of an existing subset.

17.2.1 Supported Domains

Execute; Image Object Level

17.2.2 Algorithm Parameters

Scene Name

Edit the name of the scene copy to be created.

Use Variable as Scale

Specify the scale for copying a map using a variable, rather than dening a numerical
value.
Scale

See Scale (p 127) for an explanation of the Select Scale dialog box.
Border Size

Extends the bounding box of the image object by the entered border size in pixels. (Only
available when image object level is selected as Domain.)
Include Maps With Objects Linked Via

Class lter for object link class that links objects on different maps. the subset will
contain all maps where the current object has linked objects on. (Only available when
image object level is selected as Domain.)
Exclude Other Image Objects

Dene a no-data area outside of the current image object to exclude other image objects
from further processing. (Only available when image object level is selected as Domain.)
Selecting yes brings up two further parameters:
Customize Path if set to yes, the Export Path parameter is activated

Reference Book

11 June 2014

Workspace Automation Algorithms

189

In Export Path, enter the le attributes and lepath in the form

{:Workspc.Output.Root}\:Scene.Name}_{:Variable:MyVariable}.ext or
simply C:\MyFile.tif
Desktop Export Folder

Specify the le export folder used for desktop processing. If the algorithm is run in
desktop mode, les will be stored at this location. In server processing mode, the le
location is dened in the export settings specied in the workspace.
Minimum and Maximum X, Y, Z & T Co-ordinates

Edit the co-ordinates of the subset. For the default co-ordinates orientation (0,0) in the
bottom left-hand corner, the different co-ordinates are dened as follows:

Figure 17.1. Co-ordinates of a subset

The minimum X co-ordinates describe the left border

The maximum X co-ordinates describe the right border
The minimum Y co-ordinates describe the lower border
The maximum Y co-ordinates describe the upper border.

Alternatively, click the drop-down arrow button to select available variables. Entering a
letter will open the Create Variable dialog box.
Co-ordinates Orientation

You can change the corner of the subset that is used as the calculation base for the coordinates. The default is (0,0) in the lower-left corner.
Additional Thematic Layers

Reference Book

11 June 2014

Workspace Automation Algorithms

190

17.3 Create Scene Tiles

Create a tiled copy of a scene. Each tile is a separate project with its own image layers
and thematic layers. Together the tile projects represent the complete scene as it was
before creating the tiled copy. The given co-ordinates (geocoding or pixel co-ordinates)
of the source scene of the rule set are used. Results are not included before the tiles are
processed.
After processing, you can stitch the tile results together and add them to the complete
scene within the dimensions as it was before creating the tiled copy. You can tile scenes
and subsets several times.

17.3.1 Supported Domains

Execute

17.3.2 Tile Size Parameters

Tile Height

Edit the height of the tiles to be created. Minimum height is 100 pixels.
Tile Width

Edit the width of the tiles to be created. Minimum with is 100 pixels.

17.4 Submit Scenes for Analysis

This algorithm enables you to connect subroutines with any process of the main process
tree or other subroutines. You also can also choose whether to stitch the results of the
analysis of subset copies.
A rule set that contains the submit scenes for analysis algorithm can only be executed
if you are connected to eCognition Server. Rule sets that include subroutines cannot be
processed on a local machine.

17.4.1 Supported Domains

Execute

17.4.2 Algorithm Parameters

Type of Scenes

Select the type of scene to submit to analysis the top level scenes, tiles, or subsets and
copies.
Reference Book

11 June 2014

Workspace Automation Algorithms

191

If you select top level scenes, they can only be submitted on the client, and not on a server.
This option is designed for use with actions that are used in the Analysis Builder window
and that will not be analyzed on the server.
State Filter

This eld only displays for top level scenes. The default is any state. use the drop-down
list to select a processing state: created, edited, processed, failed, canceled or rejected.
Submit Recursively

Submits scenes recursively.

Scene Name Prefix

Enter the prex of the names of scene copies to be selected for submitting. A prex is
dened as the complete scene name or the beginning of it. Enter the unique part of the
name to select only that scene, or the beginning of the name to select a group with similar
or sequential names. For example, if you have scene names 7a, 7b and 7c, you can select
them all by entering a 7, or select one by entering 7a, 7b or 7c.
Process Name

Address a subroutine or a process in the process tree of a subroutine for execution by

using a forward slash before hierarchy steps, for example, subroutine/process name. You
can use the context menu in the Process Tree window to copy a process and paste it into
this eld.
Parameter Set for Process

Select a parameter set to transfer variables to the following subroutines. Click the ellipsis
button to open the Select Parameter Set dialog box.
Percent of Tiles to Submit

If you do not want to submit all tiles for processing but only a certain percentage, you
can edit the percentage of tiles to be processed. If you change the default value of 100,
the tiles are selected randomly. If the calculated number of tiles to be submitted is not an
integer it is rounded up to the next integer.
If the value entered is less than or equal to 0, 1.0 will be used. If the value entered is
greater than 100, 100 will be used. Tiles that are not selected are automatically assigned
the status skipped.

Reference Book

11 June 2014

Workspace Automation Algorithms

192

17.4.3 Stitching Parameters

Stitch Subscenes

Select Yes to stitch the results of subscenes together and add them to the complete scene
within its original dimensions. Only the main map of a tile projects can be stitched.
Overlap Handling

If Subsets and Copies are stitched, the overlapping must be managed. You can opt to
create Intersection image objects (default) or select Union to merge the overlapping image
objects.
Class for Overlap Conflict

Overlapping image objects may have different classications. In that case, you can dene
a class to be assigned to the image objects resulting from overlap handling.
Ignore Failed Tiles or Subsets

If set to Yes tiles or subsets that have status failed will be ignored by the stitching process.
Transparency value for temporary image layers

A value to be ignored when stitching temporary image layers. Enter the value manually
or choose:
From Feature. . .
From Array Item. . .

17.4.4 Post-Processing Parameters

Request Post-Processing

Select Yes to execute another process after subscenes are processed.

Post-Process Name

Address a subroutine or a process in the process tree of a subroutine for execution by

using a forward slash before hierarchy steps, for example, subroutine/process name. You
can use the context menu in the Process Tree window to copy a process and paste it into
this eld.

Reference Book

11 June 2014

Workspace Automation Algorithms

193

Parameter Set for Post-Process

Select a parameter set to transfer variables to the following subroutines. Click the ellipsis
button to open the Select Parameter Set dialog box.

17.5 Delete Scenes

Delete the scenes you no longer want to use or store.

17.5.1 Supported Domains

Execute

17.5.2 Algorithm Parameters

Type of Subscene

Select the type of scene copy to be deleted: tiles or subsets and copies.
Scene Name Prefix

Enter the prex of the names of scene copies to be selected for deleting. A prex is
dened as the complete scene name or the beginning of it. Enter the unique part of the
name to select only that scene, or the beginning of the name to select a group with similar
or sequential names. For example, if you have scene names 7a, 7b and 7c, you can select
them all by entering a 7, or select one by entering 7a, 7b or 7c.

17.6 Read Subscene Statistics

Read exported result statistics and perform a dened mathematical summary operation.
The resulting value is stored as a variable that can be used for further calculations or
export operations concerning the main scene. This algorithm summarizes all values in
the selected column of a selected export item, using the selected summary type.
When the analysis of subscenes results in exporting statistics for each scene, the algorithm
allows you to collect and merge the statistical results of multiple les. The advantage is
that you do not need to stitch the subscenes results for result operations concerning the
main scene. The following preconditions apply:
For each subscene analysis, a project or domain statistic has already been exported
All preceding subscene analysis including export has been processed completely
before the read subscene statistics algorithm starts any result summary calculations.
To ensure this, result calculations are done within a separate subroutine.

Reference Book

11 June 2014

Workspace Automation Algorithms

194

17.6.1 Supported Domains

Execute

17.6.2 Algorithm Parameters

Use Variable as Scene Type

This function lets you use the text of a variable containing a scene type, rather than the
scene type itself. Selecting yes activates the Scene Type Variable parameter.
Scene Type Variable

Create or select a scene type variable.

Type of Scene

Select the type of scene to be submitted (this parameter only appears if Use Variable as
Scene Type is set to no):

Current scene
Subsets and copies
Subset and copies recursively
Tiles
Tiles recursively
Tiles from subset.

Scene Name Prefix

Enter the prex of the names of the scenes to be selected for reading. A prex is dened
as the complete or the beginning of the scene name. Enter the unique part of the name
to select only that scene, or the beginning of the name to select a group with similar or
sequential names. For example, if you have scene names 7a, 7b and 7c, you can select
them all by entering a 7, or select one individually by entering 7a, 7b or 7c.
Export Item

Enter the name of the export item statistics to be summarized (for example
ProjectStatistics or ObjectStatistics).
Output Type

This value species the type of statistical output:

Single Column reads single column values and writes output to the specied variable

Reference Book

11 June 2014

Workspace Automation Algorithms

195

All Columns reads all available columns, creates a variable for each one, then puts
these variables into a specied feature list
Column

If single column is selected as the Output Type, specify which column should be used
by the algorithm for the statistical summary operation.
Feature List

The Feature List parameter appears when all columns is selected as the output type.
Enter the feature list variable to receive the columns.
Add Summary Prefix

The Add Summary Prex option appears when the Output Type parameter is set to all
columns. It adds the summary type as a prex to each variables name.

17.6.3 Mathematical Parameters

The following summary operations are available:
Value

Description

Count

Calculate the number of items

Sum

Sum all values of appropriate statistics table columns

Mean

Calculate the average of all values

Std. Dev. Calculate the standard deviation of all values

Min

Calculate the minimal value of all values

Max

Calculate the maximal value of all values

The following parameters are common to all the mathematical parameters:

Use Variable for Usage Flag

If yes is selected, a variable value is used to calculate a summary; if set to no, a xed
value is used
Calculate Summary

Select yes to calculate the summary.

Reference Book

11 June 2014

Workspace Automation Algorithms

196

Result Variable

Enter a variable to store the resulting value. (This parameter is only visible if Single
Column is selected and Calculate Summary is set to yes.)

Reference Book

11 June 2014

18 Interactive Operation
Algorithms
Interactive operation algorithms are used to provide user interaction with the user of
actions in eCognition Architect.

18.1 Show User Warning

Edit and display a user warning.

18.1.1 Supported Domains

Execute

18.1.2 Algorithm Parameters

Message

Edit the text of the user warning.

18.2 Set Active Pixel

Sets the active pixel to the given coordinate.

18.2.1 Supported Domains

Execute

18.2.2 Algorithm Parameters

X Coordinate/Y Coordinate

There are two ways to set the value:

197

Interactive Operation Algorithms

198

1. Assign the value by entering a numeric value, or enter a non-numeric value to

create a variable
2. Click the drop-down list to select a feature.

18.3 Create/Modify Project

Create a new project or modify an existing one.

18.3.1 Supported Domains

Execute

18.3.2 Image Layer Parameters

Image File

Browse for an image le containing the image layers. Alternatively you can edit the path.
Image Layer ID

Change the image layer ID within the le. Note, that the ID is zero-based.
Image Layer Alias

Edit the image layer alias.

18.3.3 Thematic Layer Parameters

Thematic File

Browse for a thematic le containing the thematic layers. Alternatively, you can edit the
path.
Attribute Table File

Browse for an attribute le containing thematic layer attributes. Alternatively, you can
edit the path.
Attribute ID Column Name

Edit the name of the column of the attribute table containing the thematic layer attributes
of interest.

Reference Book

11 June 2014

Interactive Operation Algorithms

199

Thematic Layer Alias

Edit the thematic layer analysis.

18.3.4 General Settings Parameters

Show Subset Selection

Opens the Subset Selection dialog box when executed interactively.

Enable Geocoding

Activate to select the bounding coordinates based on the respective geographical coordinate system.

18.4 Manual Classification

This algorithm allows the user of an action to classify image objects of the selected class
manually by clicking on them.

18.4.1 Supported Domains

Execute

18.4.2 Algorithm Parameters

Class

Select a class that can be assigned manually.

Use Brush

The Use Brush parameter, when assigned the value Yes, activates a brush tool. Holding
down the left mouse button allows a user to manually classify objects by dragging the
mouse over them. If this action is performed while pressing the Shift key, objects are
unclassied.
Brush Size

Dene the brush size used for classication.

Reference Book

11 June 2014

Interactive Operation Algorithms

200

18.5 Configure Object Table

Display a list of all image objects together with selected feature values in the Image
Object Table window.

18.5.1 Supported Domains

Execute

18.5.2 Algorithm Parameters

Classes

Select classes to list all of its image objects.

Features

Select the features to display the feature values of the image objects.

18.6 Select Input Mode

Set the mode for user input via graphical user interface.

18.6.1 Supported Domains

Execute

18.6.2 Algorithm Parameters

Input Mode

Select an input mode:

If Normal is selected, normal input mode is returned to (for example, selection of
image objects by clicking them)
Manual Object Cut activates the Cut Objects Manually function.

18.7 Start Thematic Edit Mode

The Start Thematic Edit Mode algorithm activates thematic editing, creation of thematic
layers and drawing functions. This mode can call on any map, not just the active one.
The following actions are available after execution:

Reference Book

11 June 2014

Interactive Operation Algorithms

201

Click and hold the left mouse button as you drag the cursor across the image to
create a path with points in the image
To create points at closer intervals, drag the cursor more slowly or hold the Ctrl key
while dragging.
Release the mouse button to automatically close the polygon
Click along a path in the image to create points at each click. To close the polygon,
double-click or select Close Polygon in the context menu
To delete the last point before the polygon is complete, select Delete Last Point in
the context menu.

18.7.1 Supported Domains

Execute

18.7.2 Algorithm Parameters

Mode

Select Full Editing or Classify Only. The Classify Only option enables the editing of
class and layer name parameters only.
Class

Selct an annotation class.

Use Magnetic Snap

When this function is activated and the user draws a line around an image object with the
mouse, the line will automatically snap to the area of highest contrast.
Magnetic Snap Radius

Insert a value to specify a detection radius; the magnetic snap detects edges only within
the specied distance. A value of zero means no magnetic snap.
Thematic Layer Name

The thematic layer name for manual annotation.

18.8 Select Thematic Objects

Use the Select Thematic Objects algorithm to enable selection of thematic objects in
the user interface. The algorithm activates thematic editing and enables cursor selection
mode. It is designed to be used with actions.

Reference Book

11 June 2014

Interactive Operation Algorithms

202

18.8.1 Algorithm Parameters

Layer Name

Enter the name of the layer where thematic objects are to be selected.
Selection Mode

Choose the type of selection:

Single: enables selection of single polygons.

Polygon: enables selection of all shapes within a user-drawn polygon.
Line: enables selection of all shapes crossed by a user-drawn line.
Rectangle: enables selection of all shapes within a user-drawn rectangle.

18.9 Finish Thematic Edit Mode

Use the Finish Thematic Edit Mode algorithm to switch back from thematic editing to
image object editing and save the shapele. It is designed to be used with actions and can
be used on multi-scene projects.

18.9.1 Supported Domains

Execute

18.9.2 Algorithm Parameters

Save Changes

Selct Yes to save manual editing changes into the le.

File Name

Dene the le name of the shapes le.

18.10 Select Image Object

Select an image object designated in the image object domain in the active view. The process has the same effect as if the user manually selects the image object with the mouse.
You can use this algorithm in Architect solutions if you want to highlight an image object
or automatically display its properties in the Image Object Information window. Only the
rst image object in the selected domain will be selected.

Reference Book

11 June 2014

Interactive Operation Algorithms

203

18.10.1 Supported Domains

Image Object Level; Current Image Object; Super Object; Sub Objects; Linked Objects

18.10.2 Algorithm Parameters

None

18.11 Polygon Cut

The Polygon Cut algorithm allows users to manually cut image objects. When activated,
the client will enter polygon input mode, which allows line or polygon editing as specied
by the rule set writer. Line and polygon mode are identical except in polygon mode, the
user can close a shape by double-clicking or releasing the mouse button near the start
point. New objects can be classied according to classication parameters and there is
an option to store them in an object list variable.
When creating a polygon, displaying the context menu by right-clicking offers the following options for a open polygon:
Cancel Split deletes the entire polygon
Close and Split joins the rst and last points together
Delete Last Point removes the last point created
The following restrictions apply:
In polygon mode, if the area of the bounding box exceeds 2,048 x 2,048, a warning
will appear that the process will take some time to excecute and prompts users if
they want to continue
In polygon mode, if the area of the bounding box exceeds 10,000 x 10,000, a
warning is shown that this value cannot be exceeded and the cutting action is not
executed
In 3D and 4D projects, the cutting is always performed within a single xy slice at a
single time point.

18.11.1 Supported Domain

Execute

18.11.2 Algorithm Parameters

Mode

Polygon Cut in polygon mode, a closed polygon is drawn and the area inside it is
cut
Line Cut in line mode, a line is drawn and rasterized by creating pixel width
image objects along the line

Reference Book

11 June 2014

Interactive Operation Algorithms

204

Class

Specify the class of objects to be annotated.

Use Magnetic Snap

Select yes or no. If yes is selected, the drawn line will automatically snap to highcontrast areas, when the user drags the mouse while holding down the left mouse button
Magnetic Snap Radius

Insert a value to specify a detection radius; the magnetic snap detects edges only within
the specied distance. A value of zero means no magnetic snap. The default value is 30.
Class Filter

Specify the class of objects to be cut.

Image Object List

If selected, the image object list variable that will receive the cut image objects.
Callback Process

The path to the process invoked after each cut operation.

Preserve Object Type for Cut Objects

Select yes or no. If yes is selected, the image object type will be preserved for cut
image objects. If no is chosen, cut image objects will be marked as disconnected (this
option is less processor intensive).
Ensure Connected for Objects Created by the Cut

Select yes or no. If yes is selected, the resulting image objects will be converted to
connected 2D. If no is chosen, resulting image objects will be marked as disconnected.

18.12 Save/Restore View Settings

Save and restore view settings of the active view. Use this algorithm to congure the current view to a predened setting used in eCognition Architect solutions. This is achieved
in two steps: rst the current view settings are stored using the Save View Settings mode.
Then the active view can be restored to the saved view settings by invoking it in the
Restore View Settings mode.

Reference Book

11 June 2014

Interactive Operation Algorithms

205

18.12.1 Supported Domain

Execute

18.12.2 Algorithm Parameters

Mode

Save view settings in this mode the view settings can be saved. They will be
stored together with the algorithm. Use the View Settings parameter to save the
current view.
Restore view settings in this mode the view settings are restored to the state
represented by another process using this algorithm in the Save View Settings mode.
Use the Process Path parameter to specify the process that stores the view settings.
View Settings

This parameter is only available in the Save View Settings mode. Click the ellipsis in the
Value column (which displays Click to Capture View Settings) to store the current view
settings. In addition to main maps, the view settings for maps and window splits are also
saved.
Process Path

This parameter is only available in the Restore View Settings mode. Refer to another
process using this algorithm in Save View Settings mode to restore the view settings that
are stored in this algorithm. To refer to a process, right-click on it and select Go To to
open the Go To Process dialog box.

18.13 Display Map

This algorithm displays a specic map in the pane view. It can be used to congure the
correct view for the user in interactive Architect solutions that use multiple maps.

18.13.1 Supported Domains

Execute

18.13.2 Algorithm Parameters

Pane

Select on which pane the map should be displayed. The following values are available:
Active

Reference Book

11 June 2014

Interactive Operation Algorithms

206

First
Second
Third
Fourth
Apply to all

Map Name

Select the name of the map to be displayed in the pane view.

18.14 Define View Layout

In eCognition Architect you can dene a particular window layout for end users. The
Dene View Layout algorithm allows you to display windows split into one, two or four
with a different map displayed in each pane or a 3D projection.
The following view properties apply:
If a map is not available, the main map is displayed
When changing layouts using the same map, the zoom properties of the active
window are used
When changing maps, the image is adjusted to the size of the window
Any layout settings, such as numbers of layers or outlines, are reset when changing
layouts.

18.14.1 Supported Domains

Execute

18.14.2 Algorithm Parameters

Split Type

No Split
Split Vertically
Split Horizontally
4 Panes
Comparison View for a 2D project, two vertical panes; for a 3D project, comparisons of XY, XZ or YZ views

View Type

Independent View
Side by Side View
Swipe View

Reference Book

11 June 2014

Interactive Operation Algorithms

207

Synchronize Views

Dene how image views are laid out and what content is displayed, for each image view.
Select yes to synchronize all the map views with each other.

18.14.3 Pane Parameters

Dependent on the selected split type the following panes are available:

First
Second
Third
Fourth

Map Name

Select the name of the map to display in the respective pane.

Projection

XY, XZ, YZ or 3D

18.15 Set Custom View Settings

This algorithm allows you to dene and apply specic view settings to your project.

18.15.1 Supported Domains

Execute

18.15.2 Algorithm Parameters

Pane

Select on which pane the view settings should be applied. The following options are
available:

Active
First
Second
Third
Fourth
Apply to all

Reference Book

11 June 2014

Interactive Operation Algorithms

208

Layer

Select to which Layer the view settings should be applied. Dependent on the project and
its available data the following selections can be made:

No Change
Image Data
Samples
temporary Layer
Thematic Layer

Mode

Select one of the available View Settings modes:

No Change
Layer
Classication
Samples
From Feature. . .
From Array Item. . .

Image Data

Select one of the available Image Data settings:

No Change
Pixel View
Object Mean View
Outlines

Select which outline display settings should be applied. Available options are:

No Change
None
Opaque
Transparent

Class Transparency

Select the desired opaqueness (in %) by either entering a certain value or one of the
following options:

No Change
Value
From Feature. . .
From Array Item. . .

Reference Book

11 June 2014

Interactive Operation Algorithms

209

Zoom

Select the desired zoom factor (in %) by either entering a certain value or one of the
following options:

No Change
Value
Fit to Window
From Feature. . .
From Array Item. . .

Polygons

Select one of the following options to display the outlines of polygons:

No Change
Off
Raster
Smoothed

Map Name

Select the map to be displayed:

No Change
Map name
main
From Parent

Image object level

Select the image object level name to be displayed.

No Change
Level Name
Time Frame

Select a specic time frame to be displayed.

No Change
Value
From Feature. . .
From Array Item. . .

Image Layer
Image Layer Mixing

No Change

Reference Book

11 June 2014

Interactive Operation Algorithms

210

Set Image Layer Mixing

When selecting Set Image Layer Mixing the follwing options are available:
Select the image layer that should be visualized in red.

Image Layer Mixing - Red

Select the image layer that should be visualized in green.

Image Layer Mixing - Green

Image Layer Mixing - Blue

Select the image layer that should be visualized in blue.

Thematic Layer
Thematic Vector Objects

Select one of the following display settings for vector objects:

No Change
Off
Manual editing
View

When selecting Manual editing or View a thematic

layer can be specied where the settings should be applied.
Thematic Vector Layer Name

Thematic Layer Mixing

Select one of the following options:

No change
Set Thematic Layer Mixing
When selecting Set Thematic Layer Mixing the following thematic layer visualization properties can be selected:
Layers

Show
Outline Color
Fill color and
Tranparency

Thematic Layer Transparency If you want to select one transparency value for all selected
thematic layers insert a transparency value here in a range of 0100. This parameter is
available for Thematic Layer Mixing > Set Thematic Layer Mixing.

18.16 Change Visible Layers

Many algorithms use temporary layers for computation purposes; however, this may be
confusing for people who are running solutions in eCognition Architect. The Change
Visible Layers algorithm allows rule-set developers to specify the visibility of new image
layers. Hidden layers are hidden by default in eCognition Developer 9.0, although it is

Reference Book

11 June 2014

Interactive Operation Algorithms

211

possible to turn this option on and off in Tools > Options (for more details consult the
user guide).

18.16.1 Supported Domains

Execute

18.16.2 Algorithm Parameters

Apply to

Select whether to apply the action to image layers or thematic layers.

Mode

The following options are available:

Show single layer

Hide single layer
Show all layers
Hide all layers

Image Layer/Thematic Layer

Select the image layer or thematic layer (depending on your selection in the Apply To
eld) to show or hide.

18.17 Change Visible Map

The Change Visible Map algorithm lets you show or hide maps in image view. You can
dene which maps can be displayed.

18.17.1 Supported Domains

Execute

18.17.2 Algorithm Parameters

Show Map

Choose from the following values:

Main
From Parent

Reference Book

11 June 2014

Interactive Operation Algorithms

212

Active Map
<Create New Variable>
Hide Map

Values for the Hide Map parameter are the same as those for Show Map.
Allow Empty Image View

When Allow Empty Image View has a value of yes:

If Show Map has no value and a map is selected in the Hide Map eld, this map is
removed and a black image view is displayed
If a map is selected in the Show Map eld but Hide Map has no value, all black
image views will display the selected map

18.18 Show Slide

The Show Slide algorithm allows the user to navigate through successive slides using
next, previous and rst commands. This value is specic to workspaces.

18.18.1 Algorithm Parameters

Show Slide

Select with slide you want to be shown (next, previous or rst).

18.19 Ask Question

18.19.1 Supported Domains
Execute

18.19.2 Algorithm Parameters

Message

Amend this value to change the title of the dialog box.

Result

Insert a value that will store the result. If the user selects yes, the value is 1; if no is
selected, the value is 0.

Reference Book

11 June 2014

Interactive Operation Algorithms

213

18.20 Set Project State

The standard conguration of eCognition Developer 9.0 is to prompt the user to save
a project upon closing, even if the user has not made any changes to the le. The Set
Project State algorithm allows rule-set writers to disable this prompt.

18.20.1 Supported Domains

Execute

18.20.2 Algorithm Parameters

Mark Project As

Set this value to modied or unmodied. If modied is selected, the user will be prompted
to save; if unmodied is selected, this option will not appear.

18.21 Save/Restore Project State

This algorithm lets you save or restore the state of all maps, levels and image projects as
part of a rule set.

18.21.1 Supported Domains

Execute

18.21.2 Algorithm Parameters

Mode

Choose either save state or restore:

In save mode, the algorithm saves all maps to a temporary project le
in restore mode, the algorithm restores the previously saved state from the project
le.
Name of the State to be Saved or Restored

Enter a name in quotes, or use the drop-down list to select a variable, feature or array
item.

Reference Book

11 June 2014

Interactive Operation Algorithms

214

18.22 Show HTML Help

The Show HTML Help algorithm allows developers to incorporate an HTML le into a
rule set. This can be displayed as a pop-up window that launches in response to a user
action. This is useful for adding additional help information to a function.
The help dialog contains a checkbox with the label do not show again, which disables
the function for a period of two weeks.

18.22.1 Supported Domains

Execute

18.22.2 Algorithm Parameters

Help File Name

Dene the name of the help le. The HTML le must be created independently and
conform to the dened address.
Shapes File

Enter the name of the shape le

18.23 Configure Image Equalization

The Congure Image Equalization algorithm sets the image layer equalization display for
selected image views.

18.23.1 Supported Domains

Execute

18.23.2 Algorithm Parameters

Pane

Select on which pane the settings should be applied. The following options are available:

Active
First
Second
Third
Fourth
Apply to all

Reference Book

11 June 2014

Interactive Operation Algorithms

215

Equalization

Select one of the following equalization modes:

None
Linear
Standard Deviation
Gamma Correction
Histogram
Manual

Equalization factor

Default Value
From Feature. . .
From Array Item. . .
If you select Default Value for the Equalization factor the following values are applied:
Linear = 1%
Standard Deviation = 3.00
Gamma Correction = 0.50
Manual image equalization

If you select Manual for Equalization the following options are available:
If yes is selected, the software will expect a scene variable
of the string type as an equalization parameter. If no is selected, equalization has to be
specied explicitly.
Use variable for equalization

Equalization

Select one of the following equalization modes:

Linear
Linear Inverse
Gamma Correction (Positive)
Gamma Correction (Negative)
Gamma Correction (Positive) Inverse
Gamma Correction (Negative) Inverse

Range Type There are two ways to dene the range Min/Max or Center/Width. The
value range will not exceed the limitations of the channel; for example, the value for an
8-bit channel will always be between 0 and 255.

min = center

width
width
, max = center +
2
2

Min/Max Dependent on the selection of the Range Type you can specify the minimum
and maximum value for the layer equalization range.

Reference Book

11 June 2014

Interactive Operation Algorithms

216

Center/Width Dependent on the selection of the Range Type you can specify the center
and width values for the layer equalization range.

Apply to all Visible Layers

Select yes to apply manual equalization to all visible image

layers.
The value of the Interactive Range Editing Step represents
the step-size in pixels, for the interactive mouse action.
Interactive Range Editing Step

18.24 Create/Update Class

This algorithms allows you to create and update classes using variables or features.

18.24.1 Supported Domains

Execute

18.24.2 Algorithm Parameters

Class name

The name of the class to be created or modied. The default value <auto> will insert a
random name.
Default Value
From Feature. . .
From Array Item. . .
Superclass name

The class name of the superclass of this class.

Default Value
From Feature. . .
From Array Item. . .
Group name

The class name of the group of this class.

Default Value
From Feature. . .
From Array Item. . .

Reference Book

11 June 2014

Interactive Operation Algorithms

217

Class comment

You can insert a class comment text.

Default Value
From Feature. . .
From Array Item. . .
Class visualization red color

For the visualization of the class color select a value between [0. . . 255]. The default
value 1 selects a random red color value for this new class.
Default Value
From Feature. . .
From Array Item. . .
Class visualization green color

For the visualization of the class color select a value between [0. . . 255]. The default
value 1 selects a random green color value for this new class.
Default Value
From Feature. . .
From Array Item. . .
Class visualization blue color

For the visualization of the class color select a value between [0. . . 255]. The default
value 1 selects a random blue color value for this new class.
Default Value
From Feature. . .
From Array Item. . .
Scope

Select if the class should operate with global or local scope.

Global
Local

Reference Book

11 June 2014

19 Parameter Set Operations

Algorithms
Parameter set operations algorithms enable the automated exchange of parameter sets
among actions. Parameter sets are especially important in rule sets using workspace
automation and rule sets created for actions.

19.1 Apply Parameter Set

Writes the values stored inside a parameter set into the related variables. For each parameter in the parameter set, the algorithm scans for a variable with the same name. If
this variable exists, then the value of the variable is updated by the value specied in the
parameter set.

19.1.1 Supported Domains

Execute

19.1.2 Parameter Set Name

Select the name of a parameter set.

19.2 Update Parameter Set

Writes the values of variables into a parameter set. For each parameter in the parameter
set, the algorithm scans for a variable with the same name. If this variable exists, then the
value of the variable is written to the parameter set.

19.2.1 Supported Domains

Execute

218

Parameter Set Operations Algorithms

219

19.2.2 Parameter Set Name

Select the name of a parameter set.
TIP: Parameters are created with the Manage Parameter Sets dialog box,
which is available on the menu bar under Process or on the toolbar.

19.3 Load Parameter Set

Load a parameter set from le. Use this algorithm to exchange parameter sets among
actions.

19.3.1 Supported Domains

Execute

19.3.2 Algorithm Parameters

File Name

Dene the le name and the path of the parameter set les to be loaded. You may use the
same settings as used for a related Save Parameter Set action.
You can use the suggested le name pattern {:Workspc.OutputRoot\}
parameter_sets\paramset.psf. It denes the folder parameter_sets located
in the workspace output root folder as displayed in the Workspace Properties dialog
box. The parameter set les are named paramset, followed by the le name ending as
described below and the le extension .psf.
Click the drop-down arrow to select text elements for editing the le name pattern.
File Name Ending

Dene the le name ending to add to the File Name eld. You may use the same settings
as used for a related Save Parameter Set action. Click the drop-down arrow to select an
available variable. Alternatively, you can insert text between the quotation marks.

19.4 Save Parameter Set

Save a parameter set as a le. Use this algorithm to exchange parameter sets among
actions.

Reference Book

11 June 2014

Parameter Set Operations Algorithms

220

19.4.1 Supported Domains

Execute

19.4.2 Algorithm Parameters

Parameter Set Name

Click the drop-down arrow to select a parameter set to save to le. Alternatively, you can
enter a given parameter set name.
File Name

See the explanation contained in Load Parameter Set on the preceding page.
File Name Ending

See the explanation contained in Load Parameter Set on the previous page.

19.5 Delete Parameter Set File

Delete a parameter set le. Use this algorithm if you know that you do not need the
parameter set any more.

19.5.1 Supported Domains

Execute

19.5.2 Algorithm Parameters

File Name

See the explanation contained in Load Parameter Set on the preceding page.
File Name Ending

See the explanation contained in Load Parameter Set on the previous page.

19.6 Update Action from Parameter Set

Synchronize the values of an action according to the values of a parameter set.

Reference Book

11 June 2014

Parameter Set Operations Algorithms

221

19.6.1 Supported Domain

Execute

19.6.2 Algorithm Parameters

Action Name

Type the name of an action.

Parameter Set Name

Select the name of a parameter set.

19.7 Update Parameter Set from Action

Synchronize the values of a parameter set according to the values of an action.

19.7.1 Supported Domains

Execute

19.7.2 Algorithm Parameters

Action Name

Type the name of an action.

Parameter Set Name

Select the name of a parameter set.

19.8 Apply Active Action to Variables

Apply Active Action to Variables updates variables that have been set by action widgets.

19.8.1 Supported Domains

Execute

Reference Book

11 June 2014

Parameter Set Operations Algorithms

222

19.8.2 Algorithm Parameters

None

Reference Book

11 June 2014

20 Sample Operation
Algorithms
Use sample operation algorithms to handle samples for Nearest Neighbor classication
and to congure the Nearest Neighbor settings.

20.1 Classified Image Objects to Samples

Create a sample for each classied image object in the image object domain.

20.1.1 Supported Domains

Image Object Settings; Image Object List

20.1.2 Algorithm Parameters

None

20.2 Cleanup Redundant Samples

Remove all samples with membership values higher than the membership threshold. 1

20.2.1 Supported Domains

Image Object Level

20.2.2 Algorithm Parameters

Membership Threshold

You can modify the default value of 0.9.

1. This algorithm might produce different results each time it is executed, because the order of sample deletion is
random.

223

Sample Operation Algorithms

224

20.3 Nearest Neighbor Configuration

Select classes, features and function slopes to use for nearest neighbor classication.

20.3.1 Supported Domains

Execute

20.3.2 Algorithm Parameters

Active Classes Choose the classes you wish to use for nearest neighbor classication.
NN Feature Space

Select as many features as you like for the nearest neighbor feature space.
Function Slope

Enter the function slope for the nearest neighbor.

20.4 Delete All Samples

Delete all samples. This algorithm has no parameters.

20.4.1 Supported Domains

Execute

20.5 Delete Samples of Classes

Delete all samples of certain classes.

20.5.1 Supported Domains

Execute

20.5.2 Algorithm Parameters

Class List

Select the classes for which samples are to be deleted.

Reference Book

11 June 2014

Sample Operation Algorithms

225

20.6 Disconnect All Samples

Disconnect samples from image objects to enable creation of samples that are not lost
when image objects are deleted. Sample statistics are stored in the rule set le. This
algorithm has no parameters.
The disconnect samples algorithm needs to be applied if objects which are samples are
altered during processing. This can be the case if objects are merged for example. In
such a case the samples are lost. To allow storing the sample values in the rule set, you
need to apply the disconnect samples algorithm prior to any process which alters objects
declared as samples. The statistics will be saved in the class descriptions from your rule
set. After diconnect all samples you dont need the sample objects anymore.
If you execute Disconnect All Samples a second or several times the algorithm is not overwriting the statistics of existing disconnected samples but accumulating the new sample
information.

20.6.1 Supported Domains

Execute

20.7 Sample Selection

Use the sample selection algorithm to switch the cursor to sample selection mode using
the selected class.

20.7.1 Supported Domains

Execute

20.7.2 Algorithm Parameters

Class

Choose a class to use in selecting samples.

Use Brush

Reference Book

11 June 2014

21 Export Algorithms
Export algorithms are used to export table data, vector data and images derived from the
image analysis results. 1

21.1 Export Classification View

21.1.1 Supported Domains
Image Object Level

21.1.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Export Unclassified as Transparent

Select Yes to export unclassied image objects as transparent pixels.

1. In Windows there is a 260-character limit on lenames and lepaths (http://msdn.microsoft.com/en-us/library
/windows/desktop/aa365247%28v=vs.85%29.aspx). Trimble software does not have this restriction and can
export paths and create workspaces beyond this limitation. For examples of this feature, refer to the FAQs in
the Windows installation guide.

226

Export Algorithms

227

Enable Geo Information

Activate to add geographic information.

Default File Format

Select the export le type used for desktop processing. If the algorithm is run in desktop
mode, les will be stored in this format. In server processing mode, the le format is
dened in the export settings specied in the workspace.

21.2 Export Current View

Export the current map view to a raster le.

21.2.1 Supported Domains

Execute; Maps

21.2.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Enable Geo Information

Activate to add geo information.

Reference Book

11 June 2014

Export Algorithms

228

Save Current View Settings

Click the ellipsis button to capture current view settings. Transparency settings may affect
the appearance of the exported view. 2
Scale

See Scale (p 127) for an explanation of the Select Scale dialog box.
Default File Format

Specify how the slices of a three-dimensional scene are exported. 3 You can export the
slices into a single montage image, into multiple les or into a multi-page .tif le. To
export only a selection of slices, you can additionally change the settings under Slices
and Frames:
Select Current to export the current slice or slices of the current time frame.
Select Single to export a certain slice or slices of a certain time frame. Indicate the
slice or time frame using the slice and frame index respectively.
Select Range to export a range of slices or slices of several time frames. Indicate
the range using the slice and time frame index respectively.
The settings can be made independently. By default, both values are set to all.

21.2.3 Slices Parameters

Slice Selection

Denes which Z slices are exported.

Slice Index

Selected slice index.

2. Projects created with previous software versions will display with the current transparency settings. If you want
to use the Export Current View algorithm and preserve the current transparency settings, access the Algorithm
parameters. Then select Click to Capture Current View Settings in the Save Current View settings eld. (If you
want to preserve the original transparency settings, do not select Click to Capture Current View Settings.)
3. The Export Current View algorithm only exports the slices that lie within the selected range or time frame. That
is, if you select a range or time frame that does not contain any slices for the current map, no image le is
written.

Reference Book

11 June 2014

Export Algorithms

229

First Slice Index

First slice index.

Last Slice Index

Last slice index.

21.2.4 Frames Parameters

Frame Selection

Denes which time frames are exported.

Frame Index

Selected time frame index.

First Frame Index

First time frame index.

Last Frame Index

Last time frame index.

21.3 Export Thematic Raster Files

Export thematic raster les.

21.3.1 Supported Domains

Image Object Level

21.3.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

Reference Book

11 June 2014

Export Algorithms

230

Use Explicit Path exports to the location dened in the Export Path eld. The
default export path is {:Workspc. OutputRoot} \results\{:Item. Name}
\{:Project. Name}.v {:Project.Ver}.{:Ext}
Export Series

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Export Type

Select the type of export:

Select Image Objects to export feature values
Select Classication to export classication by unique numbers associated with
classes.
Features

Select one or multiple features to export their values.

Default File Format

Shift the geocoding lower-left corner in the X direction. Use the drop-down list to shift
half a pixel to the left or right.
Geo-Coding Shift Y

Shift the geocoding lower-left corner in the Y direction. Use the drop-down list to shift
half a pixel up or down.

21.4 Export Existing Vector Layer

Exports an already existing vector layer to a le.

21.4.1 Supported Domains

Image Object Level

Reference Book

11 June 2014

Export Algorithms

231

21.4.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Thematic Layer

Select the thematic layer to be exported.

Default File Format

Select the export le type used for desktop processing. If the algorithm is run in desktop
mode, les will stored in this format. In server processing mode, the le format is dened
in the export settings specied in the workspace.
Instead of exporting the results from multiple projects into multiple les it also possible
to export the results in multiple feature classes of one FileGDB.

21.5 Export Domain Statistics

Select an image object domain and export statistics regarding selected features to a le.

21.5.1 Supported Domains

Image Object Level; Image Object List

Reference Book

11 June 2014

Export Algorithms

232

21.5.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Domain Features

Select the features whose statistical calculations you want to be exported.

Default File Format

File Name Sufx allows users to select features or variables or enter a string. The value
of this parameter is then added to the name of the exported le.
When you dene a sufx, be aware that certain characters in Windows are invalid in
lenames invalid lename characters will result in Windows error code 123.

21.5.3 Statistical Operations

Select statistical operators with a Yes or No from the drop-down arrow. The following
six operators are available: Number, Sum, Mean, Std. Dev, Min and Max. Users can also
select in addition to Yes or No a scene variable:
0 = No
<>0 = Yes

Reference Book

11 June 2014

Export Algorithms

233

21.6 Export Project Statistics

Export values of selected project features to a le.

21.6.1 Supported Domains

Execute

21.6.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Features

Select one or multiple features for exporting their values.

Default File Format

21.7 Export Object Statistics

Export statistics of selected features per image object. Per project, one le is generated.
In spreadsheet les the following table columns are included:
inner_x: x coordinate of the inner pixel 4
4. The inner pixel is a pixel located within the image object serving as image object reference. The calculation
of inner pixel coordinates has changed with eCognition version 8 and coordinate values of inner pixel exported
with the latest version are different to those exported with earlier versions.

Reference Book

11 June 2014

Export Algorithms

234

inner_y: y coordinate of the inner pixel

level_name: The image object level of the image object
class_name: The class of the image object
An additional column per selected feature.

21.7.1 Supported Domains

Image Object Level; Image Object List

21.7.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

Selecting Yes allows you to customized the default export path.

Export Series

If set to Yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Features

Select one or multiple features for export.

Default File Format

File Name Sufx allows users to select features or variables or enter a string. The value
of this parameter is then added to the name of the exported le.

Reference Book

11 June 2014

Export Algorithms

235

21.7.3 Report Parameters

Create Summary Report

Select Yes to create a summary report after processing is complete.

Remove Auxillary Columns

Choose whether reports should contain auxillary columns.

Split File By

This feature allows you to create a separate export le based on one or more features.

21.8 Export Object Statistics for Report

Export image object statistics to a le. This generates one le per workspace.

21.8.1 Supported Domains

Image Object Level; Image Object List

21.8.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Features

Select one or multiple features for exporting their values.

Reference Book

11 June 2014

Export Algorithms

236

Default File Format

21.9 Export Vector Layer

Exports image objects to a vector le.

21.9.1 Supported Domains

Image Object Level

21.9.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.

21.9.3 Export Data Parameters

Attribute Table

Attribute table.
Shape Type

Select a type of shape for export polygons, lines or points.

Reference Book

11 June 2014

Export Algorithms

237

Export Type

Select a type of export for the type of shape selected:

Polygons (raster or smoothed)
Lines (main line or skeleton)
Points (center of main line or center of gravity)
Dimension

Dimension of vertex coordinates in the shape le. If 2D dimension for a 3D scene is

selected, then a vertical series of slices is exported.
Use geocoded Coordinates

Choose between geo coordinates (Yes) or pixels (No) for exported vertices in the shapele.
Coordinates Orientation

Species how the given coordinates should be interpreted (available for 3D data).

21.9.4 Export Format Parameters

Export Format

Click the drop-down arrow to select Shapele or GDB le format. Instead of exporting
the results from multiple projects into multiple les it also possible to export the results
in multiple feature classes of one FileGDB.
Name of Feature Class to Export

Click to open the Select Feature Class to Export dialog box and select a feature class.
This eld displays options if you use Trimble Data Management in combination with the
eCognition Server.

21.10 Export Image Object View

The Export Image Object View algorithm lets you export an image le for each image
object.

21.10.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects; Image Object List

Reference Book

11 June 2014

Export Algorithms

238

21.10.2 Output Parameters

Export Mode

Static Export Item writes the export item to the workspace

If Static Export Item or Dynamic Export Item are selected and this parameter is set to
yes, you may enter an le export location in the Export Path eld.
Export Series

If set to yes, then multiple les per series will be exported, or additional columns will
be created for table exports.
Default File Format

Enter the le name for the exported image. If multiple les are exported, this name is
used as a prex, to which numbers are added.

21.10.3 Settings
View Settings Source

Select from local or from process:

If you select local, you can use the local settings derived from the Save Current
View Settings on the following page parameter
Selecting from process displays the View Settings Process Path parameter. You
can use this parameter in conjunction with the View Settings parameter of the
Save/Restore View Settings (p 204) algorithm.

Reference Book

11 June 2014

Export Algorithms

239

Save Current View Settings

Click the ellipsis button to capture current view settings.

Border Size Around Object

Enter a value for the thickness of the pixels around the bounding box of the exported
image object (the default is 0).
Use Fixed Image Size

Select yes to export to a predetermined image size.

Size X

Enter a value to x the width ( x -direction) of the exported image.

Size Y

Enter a value to x the height ( y -direction) of the exported image.

Draw Object Outline

If yes is selected, a red outline is drawn around the image object.

21.11 Export Mask Image

Export the data of specic regions of interest to a le. A background ll color is used for
all areas of the scene copy that are not part of the selected regions of interest. For threedimensional maps, an individual le is exported per image object slice of the original
map.

21.11.1 Supported Domains

Image Object Level

21.11.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

Enter a name in Export Item Name, or use the default name

Reference Book

11 June 2014

Export Algorithms

240

Dynamic Export Item lets you enter or select a variable as an export item
Select a variable from the drop-down box or enter a name in the Export Item
Variable Name eld (entering a name launches the Create Variable dialog box,
where you can enter a value and variable type)
Use Explicit Path exports to the location dened in the Export Path eld. The
default export path is {:Workspc. OutputRoot} \results\{:Item. Name}
\{:Project. Name}.v {:Project.Ver}.{:Ext}
Export Series

If set to yes, then multiple les per series will be exported, or additional columns will be
created for table exports.
Scale

See Scale (p 127) for an explanation of the Select Scale dialog box.
Default File Format

Select the channel value of the required background ll color. You can enter any integer
value that represents an actual gray value. If you enter a value that is invalid for the
current image le, it will automatically be changed to the closest valid one.

21.12 Export Image

Export an image using predened settings.

21.12.1 Supported Domains

None

21.12.2 Algorithm Parameters

Layer

Select the layer for export.

Reference Book

11 June 2014

Export Algorithms

241

Export Mode

Static Export Item writes the export item to the workspace

Enter a name in Export Item Name, or use the default name
Dynamic Export Item lets you enter or select a variable as an export item
Select a variable from the drop-down box or enter a name in the Export Item
Variable Name eld
Use Explicit Path exports to the location dened in the Export Path eld. The
default export path is {:Workspc. OutputRoot} \results\{:Item. Name}
\{:Project. Name}.v {:Project.Ver}.{:Ext}
Default File Format

Select from the following formats if exporting in desktop mode (if you are using server
mode, then the export settings dened in the workspace will dene the le format):

ESRI ASCII GRID les (*.asc)

Tagged Image Files (Geocoded) (*.tif)
Tagged Image Files (*.tif)
eCognition Large Data File (*.tif)
Erdas Imagine Images (*.img)
JPEG JFIF (*.jpg)
JPEG2000 (*.jp2)
Portable Network Graphics (*.png)
Portable Network Graphics (8 bit)(*.png)
Windows or OS/2 Bitmaps (*.bmp)
National Imagery Transmission (NITF) (*.ntf)
Tagged Image Files (Internal)(*.tif)
LAS 1.2 (*.las)
Trident LAS 2.0 (*.las)
PCIDSK (*.pix)
16-bit png (*.png)

Geo-Coding Shift X

Shift the geocoding lower-left corner in the X direction. Use the drop-down list to shift
half a pixel to the left or right.
Geo-Coding Shift Y

Shift the geocoding lower-left corner in the Y direction. Use the drop-down list to shift
half a pixel to the left or right.

21.13 Export Result Preview

The Export Result Preview algorithm exports results as overlay images, which allows
faster previews of results.

Reference Book

11 June 2014

Export Algorithms

242

21.13.1 Supported Domains

Pixel Level; Image Object Level

21.13.2 Algorithm Parameters

Export Mode

Static Export Item writes the export item to the workspace

Select yes to customize the default export path.

Mode

Select the algorithm mode:

Create creates a new result preview layer from the domain level
Create from Subscenes creates a new result preview layer from specied subscenes
Append appends or creates a domain level into a result preview layer
Generate Overview generates an overview of an existing result preview layer.

Target Map

Enter the target map for the result preview export.

Generate Overview

Select yes to generate an overview.

Reference Book

11 June 2014

22 Image Registration
Algorithms
Image registration algorithms are used to transform two or more sets of image data into
one system of co-ordinates.

22.1 Image Registration

Perform the registration of the scenes stored in two different maps.
The resulting registered scene is stored in a new map. Landmark-based image registration
requires that you have created landmarks before. This can be done either manually using
a Trimble client or automatically by rule sets using the Set Landmarks algorithm.

Figure 22.1. Workflow for registering two 2D images

Landmarks may be created manually or automatically with results similar to A2 and B2.
B1/A1 illustrates the result of the registration of B1 with respect to A1.

243

Image Registration Algorithms

244

22.1.1 Supported Domains

Execute

22.1.2 Algorithm Parameters

Target Map Name

Name of the target map that provides the matched data.

Maps to Register

List of maps that should be registered to the map selected in the domain.
Transformation Type

Elastic (Spline) use an elastic transformation algorithm based on splines

Afne use an afne transformation.
Registration Mode

This option is only active for afne transformation types.

Landmarks use landmarks for registration
Automatic by pixel brightness use automatic registration by optimizing the matching pixel brightness.
Fixed Layer

This parameter is only available for afne transformations using the Automatic by pixel
brightness mode. Select the reference layer for pixel brightness optimization.
Iterations

This parameter is only available for afne transformations. Number of iterations when
optimizing the afne transformation parameters.
Save Transformation In Parameter Set

Name of the parameter set to save the afne transformation matrix in case of a single map
registration.

Reference Book

11 June 2014

Image Registration Algorithms

245

Resample Type

Resampling method for the transformed image layer.

Value

Description

Linear

Use linear resampling method. This methods offers a good trade-off

between quality and performance.

B-spline

Use B-spline resampling method. This method can produce better

resampling results than linear but requires more computing time.

Nearest Neighbor This resampling method is the fastest. It may produce results with less
quality than linear or B-spline.

Default Pixel Value

Pixel value to use for empty areas in the transformed image.

22.2 Delete Landmarks

Delete all landmarks in the specied map.

22.2.1 Supported Domains

Execute

22.3 Set Landmark

Set a landmark at the center of each object in the selected domain.

22.3.1 Supported Domains

Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub
Objects; Linked Objects

Reference Book

11 June 2014

23 About Features
23.1 About Features as a Source of Information
Image objects have spectral, shape, and hierarchical characteristics. These characteristic
attributes are called features in Trimble software. Features are used as source of information to dene the inclusion-or-exclusion parameters used to classify image objects.
There are two major types of features:
Object features are attributes of image objects, for example area
Global features are not connected to an individual image object, for example the
number of image objects of a certain class.

23.1.1 Conversions of Feature Values

The conversion of feature values is handled differently, depending on the value type:
Values identifying a position (position values)
Values identifying certain distance measurements such as length or are (unit values).
Conversion of Position Values

Position values can be converted from one co-ordinate system to another. The following
position conversions are available:
If the unit is a pixel, a position within the pixel co-ordinate system is identied
If the unit is a co-ordinate, a position within the user co-ordinate system is identied
The position conversion is applied for image object features such as Y center, Y max and
X center.
Conversion of Unit Values

Distance values such as length and area are initially calculated in pixels. They can be
converted to a distance unit. To convert a pixel value to a unit, the following information
is needed:
246

About Features

247

Pixel size in meters

Value dimension, for example 1 for length and 2 for area
Unit factor, relative to the meter, for example 1 for meter, 100 for centimeter and
0.001 for kilometer
The following formula is valid for converting value from pixel to a unit:
valunit = valpixel udim F
(Where u is pixel size in units, F is the unit factor and dim the dimension.)

23.2 Object Features

Object features are calculated by evaluating image objects themselves as well as their
embedding in the image object hierarchy. Object Features are grouped as follows:
Customized object features are user-dened and reference existing object features
Type features refer to the spatial connectivity of an image object
Layer value features evaluate the rst, second, and third statistical moment (mean,
standard deviation, and skewness) of an image objects pixel value and the objects
relations to other image objects pixel values. Use these to describe image objects
with information derived from their spectral properties
Geometry features evaluate the image objects shape. The basic geometry features
are calculated based on the pixels forming an image object. If image objects of
a certain class stand out because of their shape, you are likely to nd a geometry
feature that describes them
Position features refer to the position of an image object relative to the entire scene.
These features are of special interest when working with geographically referenced
data, as an image object can be described by its geographic position
Texture features are used to evaluate the texture of image objects. They include
texture features based on an analysis of sub-objects helpful for evaluating highly
textured data. In addition, features based upon the gray level co-occurrence matrix
after Haralick are available
Object Variables are local variables. In contrast to scene variables they store values
for each individual image object. (Think of it as each image object having its own
version of the variable.) There is one instance per image object in a project
Hierarchy features provide information about the embedding of an image object
within the image object hierarchy.
Thematic attribute features are used to describe an image object using information provided by thematic layers. If your scene contains a thematic layer, its thematic objects
properties, can be used to create thematic attribute features, which can be used for developing ruleware. Depending on the attributes of the thematic layer, a large range of
different features becomes available.

Reference Book

11 June 2014

24 Vector Features
Vector features allow addressing vectors by their attributes. These features are based on
vectors objects and not on image objects.

24.1 Vector object attribute

Vector Features > [Vector object attribute]
Vector features are available in the feature tree if the project includes a thematic layer,
otherwise no vector object attribute can be selected.

24.1.1 Editable Parameters

Thematic Layer Attribute

248

25 Object Features: Customized

Object Features > Customized
Customized features are Trimble features that you can create and adapt to your needs.
They can be arithmetic or relational features that depend on other existing features. All
customized features are based on the features shipped with eCognition Developer 9.0, as
well as newly created customized features.
Arithmetic features are composed of existing features, variables, and constants,
which are combined via arithmetic operations. Arithmetic features can be composed of multiple features but apply only to a single object.
Relational features are used to compare a particular feature of one object to those
of related objects of a specic class within a specied distance. Related objects are
surrounding objects such as neighbors, sub-objects, superobjects, sub-objects of a
superobject or a complete image object level. Relational features are composed of
only a single feature but refer to a group of related objects.

25.1 Create Customized Features

The Manage Customized Features dialog box allows you to add, edit, copy, and delete
customized features. It enables you to create new arithmetic and relational features based
on the existing ones.
1. To open the Manage Customized Features dialog box, do one of the following:
On the menu bar click on Tools and then select Manage Customized Features.
On the Tools toolbar click on the Manage Customized Features icon.
2. Click Add to create a new customized feature. The Customized Features dialog
box will open, providing you with tools for the creation of arithmetic and relational
features.
3. To edit a feature, select it and click Edit to open the Customized Features dialog
box
4. To copy or delete a feature, select it and click Copy or Delete.

25.2 Arithmetic Customized Features

Object Features > Customized > Create New Arithmetic Feature

249

Object Features: Customized

250

Figure 25.1. Manage Customized Features dialog box

To create an arithmetic customized feature:

1. In the Feature View window, double-click on Create New Arithmetic Feature
2. Insert a name for the customized feature to be created
3. Use the calculator to create the arithmetic expression. You can:
Type in new constants
Select features or variables in the feature tree on the right
Choose arithmetic operations or mathematical functions
4. The expression you create is displayed in the text area above the calculator
5. To calculate or delete an arithmetic expression, highlight the expression with the
cursor and then click either Calculate or Del as appropriate
6. You can switch between degrees (Deg) or radians (Rad) measurements and invert
the expression
7. To create the new feature click Apply, to create the feature without leaving the
dialog box, or OK to create the feature and close the dialog box
8. After creation, the new arithmetic feature can be found in either one of the following locations:
In the Image Object Information window
In the Feature View window under Object features > Customized
NOTE: The calculator buttons are arranged in a standard layout. In addition:

^ signies an exponent (x^2 for x2 ) or a square root (x^0.5 for x )

Use abs for an absolute value
Use oor to round down to the next lowest integer (whole value). You
can use oor (0.5+x) to round up to the next integer value.
(Note that e is the exponential function and PI (P) is .)

Reference Book

11 June 2014

Object Features: Customized

251

Figure 25.2. Creating an arithmetic feature in the Customized Features dialog box

25.3 Relational Customized Features

Object Features > Customized > Create New Relational Feature
To create a relational customized feature:
1. In the Feature View window, double-click on Create New Relational Feature
2. Insert a name for the relational feature to be created.
3. In the Relational Function area, select the relation existing between the image objects
4. In the drop-down list, choose the relational function to be applied.
5. Dene the distance of the related image objects. Depending on the related image
objects, the distance can be either horizontal (units, for example, pixels) or vertical
(image object levels)
6. In the Feature Selection pane, select the feature for which to compute the relation.
7. In Class Selection, select a class, group or no class to apply the relation.

Reference Book

11 June 2014

Object Features: Customized

252

Figure 25.3. Creating a relational feature at the Customized Features dialog box

8. To create the new feature click Apply to create the feature without leaving the
dialog box or OK to create the feature and close the dialog box
9. After creation, the new relational feature can be found in the Feature View window
under Class-Related Features > Customized.
NOTE: As with class-related features, the relations refer to the groups
hierarchy. This means if a relation refers to one class, it automatically
refers to all subclasses of this class in the groups hierarchy.

25.3.1 Relations Between Surrounding Objects

Relations between surrounding objects can exist either on the same level or on a level
lower or higher in the image object hierarchy (table 25.1, Relations Between Surrounding
Objects).

Reference Book

11 June 2014

Object Features: Customized

253

Table 25.1. Relations Between Surrounding Objects

Relation

Description

Neighbors

Related image objects on the same level. If the distance of the image
objects is set to 0 then only the direct neighbors are considered. When the
distance is greater than 0 then the relation of the objects is computed using
their centers of gravity. Only those neighbors whose center of gravity is
closer than the distance specied from the starting image object are
considered. The distance is calculated either in metric units or pixels. For
example, a direct neighbor might be ignored if its center of gravity is
further away from the specied distance.

Sub-objects

Image objects that exist under other image objects (superobjects) whose
position in the hierarchy is higher. The distance is calculated in levels.

Superobject

Contains other image objects (sub-objects) on lower levels in the

hierarchy. The distance is calculated in levels.

Sub-objects of
superobject

Only the image objects that exist under a specic superobject are
considered in this case. The distance is calculated in levels.

Level

Species the level on which an image object will be compared to all other
image objects existing at this level. The distance is calculated in levels.

25.3.2 Relational Functions

An overview of all functions existing in the drop-down list under the Relational Function
section is shown in table 25.2 on the current page, Relational Function Options.
Table 25.2. Relational Function Options

Relational
function

Description

Mean

Calculates the mean value of selected features of an image object and its
neighbors. You can select a class to apply this feature or no class if you
want to apply it to all image objects. Note that for averaging, the feature
values are weighted with the area of the image objects.

Standard
deviation

Calculates the standard deviation of selected features of an image object

and its neighbors. You can select a class to apply this feature or no class if
you want to apply it to all image objects.

Mean difference

Calculates the mean difference between the feature value of an image

object and its neighbors of a selected class. Note that the feature values
are weighted by either by the border length (distance = 0) or by the area
(distance > 0) of the respective image objects.
Continues. . .

Reference Book

11 June 2014

Object Features: Customized

254

Relational
function

Description

Mean absolute
difference

Calculates the mean absolute difference between the feature value of an

image object and its neighbors of a selected class. Note that the feature
values are weighted by either by the border length (distance = 0) or by the
area (distance > 0)of the respective image objects.

Ratio

Calculates the proportion between the feature value of an image object

and the mean feature value of its neighbors of a selected class. Note that
for averaging the feature values are weighted with the area of the
corresponding image objects.

Sum

Calculates the sum of the feature values of the neighbors of a selected

class.

Number

Calculates the number of neighbors of a selected class. You must select a

feature in order for this feature to apply, but it does not matter which
feature you pick.

Min

Returns the minimum value of the feature values of an image object and
its neighbors of a selected class.

Max

Returns the minimum value of the feature values of an image object and
its neighbors of a selected class.

Mean difference
to higher values

Calculates the mean difference between the feature value of an image

object and the feature values of its neighbors of a selected class, which
have higher values than the image object itself. Note that the feature
values are weighted by either by the border length (distance = 0) or by the
area (distance > 0) of the respective image objects.

Mean difference
to lower values

Calculates the mean difference between the feature value of an image

object and the feature values of its neighbors of a selected class, which
have lower values than the object itself. Note that the feature values are
weighted by either by the border length (distance = 0) or by the area
(distance > 0) of the respective image objects.

Portion of higher
value area

Calculates the portion of the area of the neighbors of a selected class,

which have higher values for the specied feature than the object itself to
the area of all neighbors of the selected class.

Portion of lower
value area

Calculates the portion of the area of the neighbors of a selected class,

which have lower values for the specied feature than the object itself to
the area of all neighbors of the selected class.

Portion of higher
values

Calculates the feature value difference between an image object and its
neighbors of a selected class with higher feature values than the object
itself divided by the difference of the image object and all its neighbors of
the selected class. Note that the features are weighted with the area of the
corresponding image objects.
Continues. . .

Reference Book

11 June 2014

Object Features: Customized

255

Relational
function

Description

Portion of lower
values

Calculates the feature value difference between an image object and its
neighbors of a selected class with lower feature values than the object
itself divided by the difference of the image object and all its neighbors of
the selected class. Note that the features are weighted with the area of the
corresponding image object.

Mean absolute
difference to
neighbors

Available only if sub-objects is selected for Relational function

concerning. Calculates the mean absolute difference between the feature
value of sub-objects of an object and the feature values of a selected class.
Note that the feature values are weighted by either by the border length
(distance = 0) or by the area (distance > 0) of the respective image objects.

25.4 Finding Customized Features

You can nd customized features at different places in the feature tree, depending on the
features to which they refer. For example, a customized feature depending on an object
feature is sorted below the group Object Features > Customized.
If a customized feature refers to different feature types, they are sorted in the feature
tree according to the interdependencies of the features used. For example, a customized
feature with an object feature and a class-related feature displays below Class-Related
Features.

Reference Book

11 June 2014

26 Object Features: Type

Object Features > Type
Type features are based on whether image objects are physically connected within a
scene.

26.1 Is 3D
Object Features > Type > Is 3D
The Is 3D feature checks if an image object has three dimensions. If the image object
type is 3D, the feature value is 1 (true), if not it is 0 (false).

26.2 Is Connected
Object Features > Type > Is Connected
The Is Connected feature checks if the image object is connected. If the image object
type is connected in two-dimensional space, the feature value is 1 (true), otherwise it is 0
(false).

256

27 Object Features: Layer

Values
Object Features > Layer Values
Layer value features evaluate the rst (mean), second (standard deviation), and third
(skewness) statistical moments of an image objects pixel value and the objects relations
to other image objects pixel values. Use these to describe image objects with information
derived from their spectral properties.

27.1 Mean
Object Features > Layer Values > Mean
Features in this group refer to the mean layer intensity value of an image object.

27.1.1 Brightness
Object Features > Layer Values > Mean > Brightness
Editable Parameter

To set which image layers providing the spectral information are used for calculation,
select Classication > Advanced Settings > Select Image Layers for Brightness from the
main menu. The Dene Brightness dialog box opens.
Parameters

wBk is the brightness weight of image layer k with wBk = {01

K is the number of image layers k used for calculation
wB is the sum of brightness weights of all image layers k used for calculation with
wB = Kk=1 wBk
ck (v) is mean intensity of image layer k of image object v
cmin
is the darkest possible intensity value of image layer k
k
is the brightest possible intensity value of image layer k.
cmin
k

257

Object Features: Layer Values

258

Figure 27.1. Define Brightness dialog box

Expression

c(v)
=

1 K B
wk ck (v)
wB k=1

Feature Value Range

max
[cmin
k , ck ]

Conditions

The scene includes more than one image layer

Because combined negative and positive data values would create an erroneous
feature value for brightness, it is only calculated using image layers with positive
values.

27.1.2 Layer 1/2/3

Object Features > Layer Values > Mean > Layer
The mean intensity of all pixel/voxels forming an image object.
Parameters

Pv is the set of pixels/voxels of an image object v with Pv = {(x, y, z,t) : (x, y, z,t)
v}
#Pv is the total number of pixel/voxels contained in Pv
ck (x, y, z,t) is the image layer intensity value at pixel/voxel (x, y, z,t)
is the darkest possible intensity value of image layer k
cmin
k
cmax
is the brightest possible intensity value of image layer k
k
ck is the mean intensity of image layer k.

Reference Book

11 June 2014

Object Features: Layer Values

259

Expression

ck (v) = ck (Pv ) =

1
ck (x, y, z,t)
#Pv (x,y,z,t)P
v

Feature Value
max
[cmin
k , ck ]

27.1.3 Max. Diff.

Object Features > Layer Values > Mean > Max. Diff.
Parameters

i, j are image layers

c(v)
is the brightness of image object v
ci (v) is the mean intensity of image layer i of image object v
c j (v) is the mean intensity of image layer j of image object v
cmax
is the brightest possible intensity value of image layer k
k
KB are image layers of positive brightness weight with KB = {k K : wk = 1},
where wk is the image layer weight.

Expression

max ci (v) c j (v)

i, jKB

c(v)

Feature Value Range

1
0, cmax
KB k
Typically, the feature values are between 0 and 1.
Conditions

The scene includes more than one image layer

If c(v)
= 0 the expression is undened.

27.2 Quantile
Object Features > Layer Values > Pixel Based > Quantile
Gives back the dened quantile of a dened layer.

Reference Book

11 June 2014

Object Features: Layer Values

260

Calculation: rst values are sorted from small to large. Associates the ordered values with
equally spaced sample fractions. It reads out the value at the dened quantile position.

27.2.1 Editable Parameters

Select a layer from the Layer drop-down box, from which to extract the quantile
Insert a percentage value into the Quantile eld (or get the value from a feature or
array item).

27.2.2 Parameters
The quantile is given by the element of the list Y with index equal to int(r)

Y is the list of pixel intensities (sorted in increasing order)

N is the number of pixels
p is a number between 0 and 100 (i.e. a percentage)
p
N
The rank, r = 100

27.2.3 Feature Value Range

[min layer value, max layer value]

27.3 Mode
Object Features > Layer Values > Pixel Based > Mode
Gives back the pixel value, which repeats most often per object Example: An object has
the following pixel values: 1 2222 333 444444 55 4 would be the mode, because it
appears most often
There is the possibility, that values appears equally often. Like: 11 2222 3333 4444 55,
then you can choose from the parameters, whether you want to get back the minimum
( smallest value, here: 4) the maximum ( largest value, here: 5) or the median (here:
3)

27.3.1 Editable Parameters

Image Layer; Return result Choose one of the following calculations for Return result
Minimum
Maximum
Median

Reference Book

11 June 2014

Object Features: Layer Values

261

27.4 Standard Deviation

Object Features > Layer Values > Standard Deviation
The standard deviation is calculated from the image layer intensity values of all
pixel/voxels forming an image object.

27.4.1 Layer 1/2/3

Object Features > Layer Values > Standard Deviation > Layer. . .
For each image layer, a separate standard deviation feature is listed in Feature View.
Parameters

k (v) is the standard deviation of intensity values of image layer k of all

pixel/voxels forming an image object v
Pv is the set of pixel/voxels of an image object v
#Pv is the total number of pixel/voxels contained in Pv
(x, y, z,t) are the pixel/voxel co-ordinates
ck (x, y, z,t) is the image layer intensity value at pixel/voxel (x, y, z,t)
range
range
ck
is the data range of image layer k with ck
= cmax
cmin
k .
k
Expression

k (v) = k (Pv ) =
v

u
u
1
u 1
t
c2k (x, y, z,t)

#Pv (x,y,z,t)P
#Pv
v

(x,y,z,t)Pv

!2
ck (x, y, z,t)

Feature Value Range

1 range
0, ck
2

27.5 Skewness
Object Features > Layer Values > Skewness
The Skewness feature describes the distribution of all the image layer intensity values of
all pixel/voxels that form an image object; this distribution is typically Gaussian. The
value is calculated by the asymmetry of the distribution of image layer intensity values in
an image object.
A normal distribution has a skewness of zero. A negative skewness value indicates that
an image object has more pixel/voxels with an image layer intensity value smaller than
the mean; a positive value indicates a value larger than the mean.

Reference Book

11 June 2014

Object Features: Layer Values

262

27.5.1 Layer Values

For each image layer, a separate skewness feature is listed in Feature View.
Parameters

k (v) is the skewness of intensity values of image layer k of an image object v

Pv is the set of pixel/voxels of image object v
#Pv is the total number of pixel/voxels contained in Pv
(x, y, z,t) are the pixel/voxel co-ordinates
ck (x, y, z,t) is the image layer value at pixel/voxel (x, y, z,t)
ck (v) is the mean intensity of image layer k of an image object v
cmax
is the brightest image layer intensity value of image layer k.
k

Expression

k (v) = k (Pv ) =

#Pv =

(x,y,z,t)Pv

3
ck (x, y, z,t) ck (v)

2
ck (x, y, z,t) ck (v)

!3
2

Feature Value Range

3
max 3
[(cmax
k ) , (ck ) ]

27.6 Pixel Based

Object Features > Layer Values > Pixel Based
Features in this group refer to the values of specic pixel/voxels in an image object.

27.6.1 Ratio
Object Features > Layer Values > Pixel Based > Ratio
The amount that a given image layer contributes to the total brightness.
Editable Parameters

Image Layer
Parameters

wBk is the brightness weight of image layer k

ck (v) is the mean intensity of image layer k of an image object v
c(v)
is brightness.

Reference Book

11 June 2014

Object Features: Layer Values

263

Expression

If wBk = 1 and c(v) 6= 0 then:

ck (v)
n

1ck (v)
k=1

If wBk = 0 or c(v) = 0 then the ratio is equal to 0.

Feature Value Range

[0, 1]
Conditions

The scene includes more than one image layer

Only image layers containing spectral information can be used to achieve reasonable results
Because combined negative and positive data values would create an erroneous
feature value for brightness, it is only calculated with image layers of positive
values.

27.6.2 Min. Pixel Value

Object Features > Layer Values > Pixel Based > Min. Pixel Value
The value of the pixel/voxel with the minimum layer intensity value in the image object.
Editable Parameters

Image Layer
Parameters

(x, y) are pixel/voxel co-ordinates

ck (x, y) is the image layer intensity value at pixel/voxel (x, y, z,t)
darkest possible intensity value of image layer k
cmin
k
cmax
brightest possible intensity value of image layer k
k
Pv is the set of pixel/voxels of an image object v.

Expression

min ck (x, y)

(x,y)Pv

Feature Value Range

max
[cmin
k , ck ]

Reference Book

11 June 2014

Object Features: Layer Values

264

Figure 27.2. Minimum pixel value of an image object consisting of three pixels

Conditions

None

27.6.3 Max. Pixel Value

Object Features > Layer Values > Pixel Based > Max. Pixel Value
The value of the pixel/voxel with the maximum layer intensity value of the image object.
Editable Parameters

Image Layer
Parameters

(x, y, z,t) are pixel/voxel co-ordinates

ck (x, y, z,t) are image layer intensity value at pixel/voxel (x, y, z,t)
is the darkest possible intensity value of image layer k
cmin
k
cmax
is the brightest possible intensity value of image layer k
k
Pv is the set of pixel/voxels of image object v.

Expression

max ck (x, y)

(x,y)Pv

Figure 27.3. Maximum pixel value of an image object consisting of three pixels

Reference Book

11 June 2014

Object Features: Layer Values

265

Feature Value Range

max
[cmin
k , ck ]

27.6.4 Mean of Inner Border

Object Features > Layer Values > Pixel Based > Mean of Inner Border
The mean layer intensity value of the pixel/voxels belonging to an image object, which
shares its border with other image objects, thereby forming an inner border.
Editable Parameters

Image Layer
Parameters

ck is the mean intensity of image layer k

PvInner is the set of inner border pixel/voxels of image object v with PvInner =
{(x, y, z) Pv : (x , y , z ) N6 (x, y, z) : (x , y , z )
/ Pv }
is
the
darkest
possible
intensity
value
of
image
layer k
cmin
k
cmax
is
the
brightest
possible
intensity
value
of
image
layer k.
k
Expression

ck (PvInner )

Figure 27.4. Inner border of a image object

Feature Value Range

max
[cmin
k , ck ]

Reference Book

11 June 2014

Object Features: Layer Values

266

27.6.5 Mean of Outer Border

Object Features > Layer Values > Pixel Based > Mean of Outer Border
The mean layer intensity value of pixel/voxels not belonging to an image object of interest,
which shares its border, thereby forming the outer border of the image object.
Editable Parameter

Image Layer
Parameters

ck is the mean intensity of image layer k

PvOuter is the set of outer border pixel/voxels of image object v with PvOuter =
{(x, y, z) Pv : (x , y , z ) N6 (x, y, z) : (x , y , z )
/ Pv }
is
the
darkest
possible
intensity
value
of
image
layer k
cmin
k
cmax
is
the
brightest
possible
intensity
value
of
image
layer k.
k
Expression

ck (PvOuter )

Figure 27.5. Outer borders of a image object

Feature Value Range

max
[cmin
k , ck ]

27.6.6 Border Contrast

Object Features > Layer Values > Pixel Based > Border Contrast
The Border Contrast algorithm looks at each edge pixel p , bordering an image object in
x , y and z directions and calculates how much it contrasts with the neighboring object at
the border.
Reference Book

11 June 2014

Object Features: Layer Values

267

Editable Parameter

Image Layer.
Parameters

Let v be an image object,

Let E(v) = {set of all pixel pairs (p, q) : p is in Pvouter and q is in Pvinner and q is in
N6(p) } be the set of all pixel edges of v
For e in E(v) the pixel edge contrast in layer k is contrast(e) : ck(q) ck(p)
The border contrast is dened as the mean value of the pixel edge contrasts for all pixel
edges in E(v) :
bc =

1
ck(q) ck(p)

n [e=(p,q)
in E(v)]

with n = #E(v)

Figure 27.6. Border Contrast algorithm: Where a pixel has a single border with an image

object (1), this value is taken; where it has multiple borders (2), the different is subtracted

27.6.7 Contrast to Neighbor Pixels

Object Features > Layer Values > Pixel Based > Contrast to Neighbor Pixels
The mean difference in contrast compared to a surrounding volume of a given size. This
feature is used to nd borders and gradations in the scene.
Editable Parameters

Image Layer; Distance (edit the size of the surrounding volume)

Reference Book

11 June 2014

Object Features: Layer Values

268

Parameters

Bv (d) is the extended bounding box of an image object v with distance d with Bv (d)
equal to {(x, y, z) : xmin (v) d x xmax (v) + d , ymin (v) d y ymax (v) + d ,
zmin (v) d z zmax (v) + d}
Pv is the set of pixel/voxels of an image object v
ck is the mean intensity of image layer k.
Expression

ck Bv (d) Pv
1000 1
1 + ck (Pv )

Figure 27.7. The surrounding area of and image object v defined by its bounding box with a

distance of one pixel

Feature Value Range

[1000, 1000]
Conditions

The distance d should always be greater than 0

If d = 0 , then Bv (d) = Bv ; if Bv = Pv , the formula is invalid
If unsigned data exists then maybe ck (Pv ) = 1 the formula is invalid
If ck (Pv ) = 0 , the values are meaningless.

27.6.8 Edge Contrast of Neighbor Pixels

Object Features > Layer Values > Pixel Based > Edge Contrast of Neighbor Pixels
The Edge Contrast of Neighbor Pixels feature describes the edge contrast of an image
object to the surrounding volume of a given size. It is used to nd edges in a scene. For
calculation, the pixel/voxels in the surrounding volume of a given size are compared to
the mean image layer intensity of the image object, subdivided into groups of brighter
and darker pixel/voxels. The value is calculated by the mean of the image layer intensity
value of the brighter pixel/voxels subtracted by the mean value of the darker pixel/voxels.

Reference Book

11 June 2014

Object Features: Layer Values

269

Figure 27.8. Pixels of the surrounding area are compared to the mean image layer intensity

of the image object v

Editable Parameters

Image Layer; Distance (the size of the surrounding area)

Parameters

Pbrighter : {P : ck (x, y, z) > ck (v)} within the surrounding volume of image object v
dened by the bounding box (Bv (d))
Pdarker : (Bv (d)) = {P : ck (x, y, z) < ck (v)} within the surrounding volume of image
object v dened by the bounding box (Bv (d))
(Bv (d)) is the extended bounding box of an image object v with distance d with
{(x, y, z) : xmin (v) d x xmax (v) + d , ymin (v) d y ymax (v) + d , zmin (v)
d z zmax (v) + d}
ck (x, y, z) is the image layer intensity value at pixel/voxel (x, y, z)
ck (v) is the mean intensity of image layer k of all pixel/voxels forming an image
object v
Pv is the set of pixel/voxels of an image object v.
Expression

ck (Pbrighter ) ck (Pdarker )
Feature Value Range

[0, 255]
Conditions

If Pbrighter (Bv (d)) = or Pdarker (Bv (d)) = then ck (v) is returned.

Reference Book

11 June 2014

Object Features: Layer Values

270

27.6.9 StdDev. to Neighbor Pixels

Object Features > Layer Values > Pixel Based > Std Dev. to Neighbor Pixels
The standard deviation of layer intensity values of pixel/voxels within the surrounding
volume of a given size. The surrounding volume consists of the pixel/voxels located
within the extended bounding box of an image object, but do no belong this image object.
Editable Parameters

Image Layer; Distance (edit the size of the surrounding volume)

Parameters

Pv is the set of pixel/voxels of an image object v

(Bv (d)) is the extended bounding box of an image object v with distance d with
{(x, y, z) : xmin (v) d x xmax (v) + d , ymin (v) d y ymax (v) + d , zmin (v)
d z zmax (v) + d}.
Expression

k (Bv (d) Pv )
Feature Value Range

cmax
0, k
2

Conditions

If d = 0 , then Bv (d) = Bv and if Bv = Pv the formula is invalid.

27.6.10 Circular Mean

Object Features > Layer Values > Pixel Based > Circular Mean
Calculates the mean for all pixels within a ring around the center of an image object. The
thickness of the ring is dened by two radius values R1 and R2 .
Editable Parameters

Layer
Radius Mode
Number
q of Pixels radius calculated from the number of pixels in the object
R = #Pxl

Length radius will equal half the length

Width radius will equal half the width

Reference Book

11 June 2014

Object Features: Layer Values

271

User radius will be specied in next parameter

Delta/User denes delta in pixels for radius of number of pixels, length, width
and user modes. It can be positive, negative or zero
Second Radius Mode
Same ( = R1 ), border second radius will be equal to rst radius. Pixels will
be used that form a thin border along the rst radius
None (=0), entire circle second radius is set to zero. Entire circle of rst
radius will be processed
Relative to R1 second radius is set to R1 . Users can dene an offset (positive, negative number or zero) to where the second radius will be extended
(positive number) or shortened (negative number)
User radius will be specied in next parameter.
Parameters

R1 is the inner radius of the ring

R2 is the outer radius of the ring
c is the center of the object
d(u, v) is the distance between objects.

Expression

(R1 0.5) d(u, v) (R2 + 0.5)

Feature Value Range

[R1 , R2 ]
Conditions

None

27.6.11 Circular StdDev

Object Features > Layer Values > Pixel Based > Circular StdDev
Calculates the standard deviation for all pixels within a ring around the center of an image
object. The thickness of the ring is dened by two radius values, R1 and R2 . Parameters
and expressions are identical to the Circular Mean feature.

27.6.12 Circular Std Dev/Mean

Object Features > Layer Values > Pixel Based > Circular Std Dev/Mean
Calculates the ratio of the mean to the standard deviation for all pixels within a ring
around the center of an image object. The thickness of the ring is dened by two radius
values, R1 and R2 . Parameters and expressions are identical to the Circular Mean feature.

Reference Book

11 June 2014

Object Features: Layer Values

272

27.7 To Neighbors
Object Features > Layer Values > To Neighbors

27.7.1 Mean Diff. to Neighbors

Object Features > Layer Values > To Neighbors > Mean Diff. to Neighbors
The Mean Diff. to Neighbors feature describes the difference between an image object
and its neighbor image objects, in terms of mean layer intensity values. For each neighboring image object of the image object of interest, the mean layer difference is computed
and weighted as follows:
If the neighboring image objects are direct neighbors (feature distance = 0), the
length of the border between the image objects is used for weighting
If the neighborhood of image objects is dened within a certain perimeter around
the image object of interest (feature distance > 0), the area covered by the neighbor
image objects is used for weighting.
Editable Parameters

Image Layer
Feature Distance: Radius of the perimeter in pixel/voxel. Direct neighbors have a
value of 0.
Parameters

u, v are image objects

b(v, u) is the length of the common border between v and u
ck is the mean intensity of image layer k
is the darkest possible intensity value of image layer k
cmin
k
cmax
is the brightest possible intensity value of image layer k
k
#Pu is the total number of pixels/voxels contained in Pv
d is the distance between neighbors
w is the image layer weight with w = uNv (d) wu
b(v,u),d=0)

wu is the weight of image object u with wu = {#Pu ,d>0

Nv is the direct neighbor to image object v with Nv : {u Vi : (x, y) Pv (x , y )
Pu : (x y ) N4 (x, y)}
Nv (d) is a neighbor to v at a distance d with Nv (d) = {u Vi : d(v, u) d}.
Expression

1
wu ck (v) ck (u)
k (v) =

w uN (d)
v

Feature Value Range

max
[cmin
k , ck ]

Reference Book

11 June 2014

Object Features: Layer Values

273

Figure 27.9. Direct neighbors (x) and the neighborhood within perimeter d of image object v

Conditions

If w = 0 the feature value is 0; therefore the formula is invalid.

27.7.2 Mean Diff. to Neighbors (Abs)

Object Features > Layer Values > To Neighbors > Mean Diff. to Neighbors (Abs.)
The Mean Diff. to Neighbors (Abs) feature describes the difference between an image
object and its neighbor image objects, in terms of their mean layer intensity values. It is
calculated in a similar way to the Mean Diff. to Neighbors feature; the only difference is
that absolute values of differences are averaged.
Editable Parameters

Image Layer
Feature Distance the radius of the perimeter in pixels. Direct neighbors have a
value of 0.
Parameters

u, v are image objects

b(v, u) is the length of the common border between v and u
ck is the mean intensity of image layer k
is the darkest possible intensity value of image layer k
cmin
k
is the brightest possible intensity value of image layer k
cmax
k
range
is the data range of image layer k
ck
d is the distance between neighbors
#Pu is the total number of pixels contained in Pu
w is the image layer weight with w = uNv (d) wu
b(v,u),d=0)

wu is the weight of image object u with wu = {#Pu ,d>0

Nv is the direct neighbor to image object v with Nv : {u Vi : (x, y) Pv (x , y )
Pu : (x y ) N4 (x, y)}
Reference Book

11 June 2014

Object Features: Layer Values

274

Nv (d) is a neighbor to v at a distance d with Nv (d) = {u Vi : d(v, u) d}.

Expression

1
k (v) =
wu |ck (v) ck (u)|

w uN (d)
v

Feature Value Range

range

[0, ck

Conditions

If w = 0 the feature value is 0; therefore the formula is invalid.

27.7.3 Mean Diff. to Darker Neighbors

Object Features > Layer Values > To Neighbors > Mean Diff. to Darker Neighbors
The Mean Diff. to Darker Neighbors feature describes the difference between an image
object and its darker neighbor image objects, in terms of their mean layer intensity values.
In contrast to the Mean Diff. to Neighbors feature, only direct neighbor image objects
are counted that have a lower mean layer value than the image object of interest.
Editable Parameters

Image Layer
Parameters

u, v are image objects

b(v, u) is the length of the common border between v and u
ck is the mean intensity of image layer k
is the darkest possible intensity value of image layer k
cmin
k
cmax
is the brightest possible intensity value of image layer k
k
range
range
ck
is the data range of image layer k , with ck
= cmax
cmin
k
k
w is the image layer weight with w = uNv (d) wu , where wu is the weight of image
b(v,u),d=0)

object u with wu = {#Pu ,d>0

Nv is the direct neighbor to image object v with Nv : {u Vi : (x, y) Pv (x , y )
Pu : (x y ) N4 (x, y)}
NvD is the darker direct neighbor to v , with NvD {u Nv : ck (u) < ck (v)}.
Expression

1
D
wu ck (v) ck (u)

k (v) =
w uN D
v

Reference Book

11 June 2014

Object Features: Layer Values

275

Feature Value Range

range

[ck

range

, ck

Conditions

If w = 0 the feature value is 0; therefore the formula is invalid. If NvD = 0 the formula is
invalid.

27.7.4 Mean Diff. to Brighter Neighbors

Object Features > Layer Values > To Neighbors > Mean Diff. to Brighter Neighbors
The Mean Diff. to Brighter Neighbors feature describes the difference between an image
object and its darker neighbor image objects, in terms of their mean layer intensity values.
In contrast to the Mean Diff. to Neighbors feature, only direct neighbor image objects
are counted that have a mean layer value less than that of the image object of interest.
Editable Parameters

Image Layer
Parameters

u, v are image objects

b(v, u) is the length of the common border between v and u
ck is the mean intensity of image layer k
cmin
is the darkest possible intensity value of image layer k
k
cmax
is the brightest possible intensity value of image layer k
k
range
range
ck
is the data range of image layer k , with ck
= cmax
cmin
k
k
w is the image layer weight with w = uNv (d) wu , where wu is the weight of image
b(v,u),d=0)

object u with wu = {#Pu ,d>0

Nv is the direct neighbor to image object v with Nv : {u Vi : (x, y) Pv (x , y )
Pu : (x y ) N4 (x, y)}
NvB is the brighter direct neighbor to v , with NvB {u Nv : ck (u) < ck (v)}.
Expression

1
Bk (v) =
B wu ck (v) ck (u)
w uN
v

Feature Value Range

range

[ck

range

, ck

Reference Book

11 June 2014

Object Features: Layer Values

276

27.7.5 Number of Brighter Objects

Object Features > Layer Values > To Neighbors > Number of Brighter Objects
The Number of Brighter Objects feature counts the neighboring objects those objects
with a common border with a higher intensity. Neighboring objects with the same mean
value are not counted.

27.7.6 Number of Darker Objects

Object Features > Layer Values > To Neighbors > Number of Darker Objects
The Number of Darker Objects feature counts the neighboring objects those objects
with a common border with a lower intensity. Neighboring objects with the same mean
value are not counted.

27.7.7 Rel. Border to Brighter Neighbors

Object Features > Layer Values > To Neighbors > Rel. Border to Brighter Neighbors
The Rel. Border to Brighter Neighbors feature describes the extent to which an image
object is surrounded by brighter or darker direct neighbor image objects. In a given image
layer, it is the ratio of the shared image border of an image object to the total border. A
value of 1 that the image object is surrounded completely by brighter neighbors; a value
of 0 means it has no brighter direct neighbors.
Editable Parameters

Image Layer
Parameters

NvB is the darker direct neighbor to v , with NvB {u Nv : ck (u) < ck (v)}
bv is the image object border length
b(v, u) is the length of common border between v and u.
Expression

uNv

b(v, u)
bv

Feature Value Range

[0, 1]

Reference Book

11 June 2014

Object Features: Layer Values

277

27.8 To Superobject
Object Features > Layer Values > To Superobject

27.8.1 Mean Diff. to Superobject

Object Features > Layer Values > To Superobject > Mean Diff. to Superobject
The difference between the mean layer intensity value of an image object and the mean
layer intensity value of its superobject.
Editable Parameters

Image Layer
Image Object Level Distance the upward distance of image object levels in the
image object hierarchy between the image object and the superobject.
Parameters

ck is the mean intensity of image layer k

range
range
cmin
= cmax
is the data range of image layer k , with ck
ck
k
k
Sv (d) is the sub-object of image object v with level distance d
Uv (d) is the superobject of image object v with level distance d.

Expression

ck (v) ck Uv (d)
Feature Value Range
range

[ck

range

, ck

27.8.2 Ratio to Superobject

Object Features > Layer Values > To Superobject > Ratio to Superobject
The ratio of the mean layer intensity value of an image object and the mean layer intensity
value of its superobject.
Editable Parameters

Image Layer
Image Object Level Distance the upward distance of image object levels in the
image object hierarchy between the image object and the superobject.

Reference Book

11 June 2014

Object Features: Layer Values

278

Parameters

ck is the mean intensity of image layer k

Uv (d) is the superobject of image object v with level distance d
Expression

ck (v)

ck Uv (d)
Feature Value Range

[0, ]
Conditions

If Uv (d) = or Uv (d) = 0 the expression is undened.

27.8.3 Std. Dev. Diff. to Superobject

Object Features > Layer Values > To Superobject > Std. Dev. Diff to Superobject
Editable Parameters

Image Layer
Image Object Level Distance the upward distance of image object levels in the
image object hierarchy between the image object and the superobject.
Parameters

k (v) is the standard deviation of intensity values of image layer k of all

pixel/voxels forming an image object v
Uv (d) is the superobject of image object v with level distance d
range
range
= cmax
cmin
is the data range of image layer k with ck
ck
k .
k
Expression

k (v) k Uv (d)
Feature Value Range

1 range 1 range
[ ck , ck ]
2
2

Reference Book

11 June 2014

Object Features: Layer Values

279

Condition

If Uv (d) = the expression is undened.

27.8.4 Std. Dev. Ratio to Superobject

Object Features > Layer Values > To Superobject > Std. Dev Ratio to Superobject
The ratio of the standard deviation of the layer intensity of an image object to the standard
deviation of the layer intensity of its superobject.
Editable Parameters

Image Layer
Image Object Level Distance the upward distance of image object levels in the
image object hierarchy between the image object and the superobject.
Parameters

k (v) is the standard deviation of intensity values of image layer k of all

pixel/voxels forming an image object v
Uv (d) is the superobject of image object v with level distance d
Expression

k (v)
k (Uv (d))
Feature Value Range

[0, ]
Conditions

If Uv (d) = or Uv (d) = 0 the expression is undened.

If v (Uv (d)) = 0 the standard deviation ratio to (Uv (d)) = 1.

27.9 To Scene
Object Features > Layer Values > To Scene

27.9.1 Mean Diff. to Scene

Object Features > Layer Values > To Scene > Mean Diff. to Scene

Reference Book

11 June 2014

Object Features: Layer Values

280

Editable Parameters

Image Layer
Parameters

ck (v) is the mean intensity of image layer k of all pixels forming an image object v
ck is the mean intensity of image layer k
range
range
= cmax
cmin
is the data range of image layer k , with ck
ck
k
k
Expression

ck (v) ck
Feature Value Range
range

[ck

range

, ck

27.9.2 Ratio To Scene

Object Features > Layer Values > To Scene > Ratio to Scene
Editable Parameter

Image Layer
Parameters

ck (v) is the mean intensity of image layer k of all pixels forming an image object v
ck is the mean intensity of image layer k.
Expression

ck (v)
ck
Feature Value Range

[, ]
Condition

If ck = 0 the expression is undened.

Reference Book

11 June 2014

Object Features: Layer Values

281

27.10 Hue, Saturation, Intensity

Object Features > Layer Values > Hue, Saturation, Intensity
Hue, Saturation, and Intensity features convert RGB color space values to HSI values.

27.10.1 HSI Transformation

Editable Parameters

Layer red, layer green, layer blue for each parameter, assign a corresponding
image layer from the drop-down list. By default these are the rst three image
layers of the scene.
Output: Select the type of HSI transformation feature to be created: hue (color),
saturation or intensity (brightness).
Parameters

RGB values are expressed as numbers from 0 to 1

max is the greatest of the RGB values
min is the smallest of the RGB values.
Expression

undened

if max = min

GB
60 maxmin
360

if max = R

BR +120

60 maxmin

360

60 RG +240

maxmin
360

if max = G
if max = B

Feature Value Range

[0, 1]

Reference Book

11 June 2014

28 Object Features: Geometry

Object Features > Geometry
Geometry features are based on an image objects shape, calculated from the pixels that
form it. Because images are raster-based, geometry features may be rotation variant: after
image objects are rotated, different feature values may arise.

28.1 Extent
Object Features > Geometry > Extent

28.1.1 Area
Object Features > Geometry > Extent > Area
The number of pixels forming an image object. If unit information is available, the number of pixels can be converted into a measurement. In scenes that provide no unit information, the area of a single pixel is 1 and the area is simply the number of pixels that
form it. If the image data provides unit information, the area can be multiplied using the
appropriate factor.
Parameters

Av is the area of image object v

#Pv is the total number of pixels contained in Pv
u is the pixel size in co-ordinate system units. If the unit is a pixel, then u = 1.
Expression

Av = #Pv u2
Feature Value Range

[0, scene size]

282

Object Features: Geometry

283

28.1.2 Border Length [for 2D Image Objects]

Object Features > Geometry > Extent > Border Length
The border length of an image object is dened as the sum of edges of the image object
shared with other image objects, or situated on the edge of the entire scene. For a torus
and other image objects with holes the border length is the sum of the inner and outer
borders.
Parameters

bv is the border length of image object

bo is the length of outer border
bi is the length of inner border
Expression

bv = bo + bi

Figure 28.1. Border length of an image object v , or between two objects v and u

Figure 28.2. Inner and outer border length of a torus image object

Feature Value Range

[0, ]

Reference Book

11 June 2014

Object Features: Geometry

284

28.1.3 Border Length [for 3D Image Objects]

Object Features > Geometry > Extent > Border Length
The border length of a 3D image object is the sum of border lengths of all image object
slices, multiplied by the spatial distance between them.
Image object slices are 2D pieces of the image object in each slice. The border length
of an image object slice is dened as the sum of edges shared with other image object
pieces, or are situated on the edge of the entire slice. For a torus and other image objects
with holes the border length sums the inner and outer border.
Parameters

bv is the border length of image object v

bv (slice) is the border length of image object slice
bv (z) is the border length of image object in z -direction
uslices is the spatial distance between slices in the co-ordinate system unit
bo is the length of the outer border
bi is the length of the inner border

Expression
#(slices)

bv =

n=1

bv (slice) uslices + bv (z)

where bv (slice) = bo + bi
Feature Value Range

[0, ]

28.1.4 Length [for 2D Image Objects]

Object Features > Geometry > Extent > Length
The length of a 2D image object is calculated using the length-to-width ratio.
Parameters

#Pv is the total number of pixels contained in Pv

v is the length-width ratio of an image object v
Expression

#Pv v

Reference Book

11 June 2014

Object Features: Geometry

285

Feature Value Range

[0, ]

28.1.5 Length [for 3D Image Objects]

Object Features > Geometry > Extent > Length
The length of an image object is the largest of three eigenvalues of a rectangular 3D
space that is dened by the same volume, and same proportions of eigenvalues, as the
image object. The length of an image object can be smaller or equal than the largest of
dimensions of the smallest rectangular 3D space enclosing the image object.
Feature Value Range

[0, ]

28.1.6 Length/Thickness
Object Features > Geometry > Extent > Length/Thickness
The length-to-thickness ratio of an image object.
Parameters

Length of the image object

Thickness of the image object
Expression

Length
Thickness
Feature Value Range

[0, ]

28.1.7 Length/Width [for 2D Image Objects]

Object Features > Geometry > Extent > Length/Width
The length-to-width ratio of an image object. There are two methods to approximate this:
1. The ratio of length to width is identical to the ratio of the eigenvalues of the covariance
matrix, with the larger eigenvalue being the numerator of the fraction:
EV
v =

1 (v)
2 (v)

Reference Book

11 June 2014

Object Features: Geometry

286

2. The ratio of length to width can also be approximated using the bounding box:
2
kvbb
BB
=
v
#Pv
Both calculations are compared; the smaller of both results is returned as the feature
value.
Parameters

#Pv is the Size of a set of pixels of an image object v

1 2 are eigenvalues
EV
v is the ratio length of v of the eigenvalues
BB
v is the ratio length of v of the bounding box
v is the length-width ratio of an image object v

kvBB

hBB
v
a is the bounding box ll rate
#Pxl
h
w is the image layer weight
p
1
2

kvbb = (kvbb )2 + (1 a)(hbb

v )
a = bb#Pv bb
kv hv

k h = #Pv k =

#Pv
w ,h

#Pv
k

k
w

k2
#Pxl

#Pxl
w2

Expression
BB
v = minEV
v , maxv

Feature Value Range

[0, ]

28.1.8 Length/Width [for 3D Image Objects]

Object Features > Geometry > Extent > Length/Width
The length-to-width ratio of an image object.
Parameters

Length of the image object

Width of the image object
Expression

Length
Width
Reference Book

11 June 2014

Object Features: Geometry

287

Feature Value Range

[0, ]

28.1.9 Number of Pixels

Object Features > Geometry > Extent > Number of Pixels
The number of pixels forming an image object. Unit information is not taken into account.
Parameters

#Pv is the total number of pixels contained in Pv

Feature Value Range

[0, scene size]

28.1.10 Thickness
Object Features > Geometry > Extent > Thickness
The thickness of an image object is the smallest of three eigenvalues of a rectangular 3D
space with the same volume as the image object and the same proportions of eigenvalues
as the image object. The thickness of an image object can be smaller or equal to the
smallest of dimensions of the smallest rectangular 3D space enclosing the image object.
Feature Value Range

[0, ]

28.1.11 Volume
Object Features > Geometry > Extent > Volume
The number of voxels forming an image object rescaled by using unit information for x
and y co-ordinates and distance information between slices.
Parameters

Vv is the volume of image object v

#Pv is the total number of voxels contained in Pv
u is the size of a slice pixel in the co-ordinate system unit
uslices is the spatial distance between slices in the co-ordinate system unit

Reference Book

11 June 2014

Object Features: Geometry

288

Expression

Vv = #Pv u2 uslices
Feature Value Range

[0, scene size]

28.1.12 Width [for 2D Image Objects]

Object Features > Geometry > Extent > Width
The width of an image object is calculated using the length-to-width ratio.
Parameters

#Pv is the total number of pixels contained in Pv

v is the length/width ratio of an image object v
Expression

#Pv
v
Feature Value Range

[0, ]

28.1.13 Width [for 3D Image Objects]

Object Features > Geometry > Extent > Width
The width of an image object is the mid-point of three eigenvalues of a rectangular 3D
space with the same volume as the image object and the same proportions of eigenvalues
as the image object. The width of an image object can be smaller or equal to the mid-point
of dimensions of the smallest rectangular 3D space enclosing the image object.
Feature Value Range

[0, ]

28.2 Shape
Object Features > Geometry > Shape

Reference Book

11 June 2014

Object Features: Geometry

289

28.2.1 Asymmetry [for 2D Image Objects]

Object Features > Geometry > Shape > Asymmetry
The Asymmetry feature describes the relative length of an image object, compared to a
regular polygon. An ellipse is approximated around a given image object, which can be
expressed by the ratio of the lengths of its minor and the major axes. The feature value
increases with this asymmetry.
NOTE: We recommend that you use the length/width ratio because it is
more accurate.

Parameters

Var X is the variance of X

Var Y is the variance of Y
Expression

q
2 14 (VarX + VarY )2 + (VarXY )2 VarX VarY
VarX + VarY

Figure 28.3. Asymmetry for a 2D image object

Feature Value Range

[0, 1]

28.2.2 Asymmetry [for 3D Image Objects]

Object Features > Geometry >Shape > Asymmetry
The Asymmetry feature describes the relative length of an image object, in the same
manner as 2D image objects. The asymmetry is calculated from the ratio between the
smallest and largest eigenvalues of the image object.

Reference Book

11 June 2014

Object Features: Geometry

290

Parameters

min is the minimal eigenvalue

max is the maximal eigenvalue
Expression

min
max

Figure 28.4. Asymmetry based on maximal and minimal eigenvalues

Feature Value Range

[0, 1]

28.2.3 Border Index

Object Features > Geometry > Shape > Border Index
The Border Index feature describes how jagged an image object is; the more jagged,
the higher its border index. This feature is similar to the Shape Index feature, but the
Border Index feature uses a rectangular approximation instead of a square. The smallest
rectangle enclosing the image object is created and the border index is calculated as the
ratio between the border lengths of the image object and the smallest enclosing rectangle.
Parameters

bv is the image object border length

lv is the length of an image object v
wv is the width of an image object v
Expression

bv
2(lv + wv )

Reference Book

11 June 2014

Object Features: Geometry

291

Figure 28.5. Border index of an image object v

Feature Value Range

[1, ] ; 1 = ideal.

28.2.4 Compactness [for 2D Image Objects]

Object Features > Geometry > Shape > Compactness
The Compactness feature describes how compact an image object is. It is similar to
Border Index, but is based on area. However, the more compact an image object is, the
smaller its border appears. The compactness of an image object is the product of the
length and the width, divided by the number of pixels.
Parameters

lv is the length of an image object v

wv is the width of an image object v
#Pv is the total number of pixels contained in Pv
Expression

21 22 23
Vv

Figure 28.6. Compactness of an image object v

Reference Book

11 June 2014

Object Features: Geometry

292

Feature Value Range

[0, ] ; 1 = ideal.

28.2.5 Compactness [for 3D Image Objects]

Object Features > Geometry >Shape > Compactness
Compactness describes how compact a 3D image object is. Appropriately scaled eigenvectors of an image objects covariance matrix provide a rough gure of the objects
extent in three dimensions.
A gure for the compactness of a 3D image object is calculated by a scaled product of
its three eigenvalues (2 1 , 2 2 , 2 3 ) divided by the number of its pixel/voxel. We
include a factor of 2 with each eigenvalue, since i eigenvectors represent otherwise
half axes of an ellipsoid dened by its covariance matrix. The chosen approach therefore
provides an estimate of a cuboid occupied by the object.
Parameters

1 is eigenvalue 1 of a 3D image object v

2 is the eigenvalue 2 of a 3D image object v
3 is eigenvalue 3 of a 3D image object v
Vv is the volume of image object v

Expression

21 22 23
Vv
Feature Value Range

[0, ] ; 1 = ideal.

28.2.6 Density [for 2D Image Objects]

Object Features > Geometry > Shape > Density
The Density feature describes the distribution in space of the pixels of an image object.
In eCognition Developer 9.0 the most dense shape is a square; the more an object is
shaped like a lament, the lower its density. The density is calculated by the number of
pixels forming the image object divided by its approximated radius, based on the covariance matrix.
Parameters

#Pv is the diameter of a square object with #Pv pixels

VarX +VarY is the diameter of the ellipse

Reference Book

11 June 2014

Object Features: Geometry

293

Expression

#Pv

1 + VarX +VarY
Feature Value Range

[0, depending on shape of image object]

28.2.7 Density [for 3D Image Objects]

Object Features > Geometry > Shape > Density
Using the same principle as Density [for 2D Image Objects], the most dense shape for
a 3D object is a cube. The more lament-shaped an image object is, the lower its density.
The value is calculated by dividing the edge of the volume of a tted cuboid by the radius
of a tted sphere.
Parameters

V
v is the volume of image object v

3
Vv is the edge of the volume tted cuboid
p
Var(X) + Var(Y ) + Var(Z) is the radius of the tted sphere
Expression

3
Vv
p
Var(X) + Var(Y ) + Var(Z)
Feature Value Range

[0, depending on shape of image object]

28.2.8 Elliptic Fit [for 2D Image Objects]

Object Features > Geometry > Shape > Elliptic Fit
The Elliptic Fit feature describes how well an image object ts into an ellipse of similar
size and proportions. While 0 indicates no t, 1 indicates a perfect t.
The calculation is based on an ellipse with the same area as the selected image object. The
proportions of the ellipse are equal to the length to the width of the image object. The
area of the image object outside the ellipse is compared with the area inside the ellipse
that is not lled by the image object.

Reference Book

11 June 2014

Object Features: Geometry

294

Parameters

v (x, y) is the elliptic distance at a pixel (x, y)

Pv is the set of pixels of an image object v
#Pv is the total number of pixels contained in Pv
Expression

# (x, y)Pv : v (x, y) 1
1
= 2
#Pv

Figure 28.7. Elliptic fit of an image object v

Feature Value Range

[0, 1] ; 1 = complete tting, 0 = <50% t.

28.2.9 Elliptic Fit [for 3D Image Objects]

Object Features > Geometry > Shape > Elliptic Fit
The Elliptic Fit feature describes how well an image object ts into an ellipsoid of similar
size and proportions. While 0 indicates no t, 1 indicates a perfect t.
The calculation is based on an ellipsoid with the same volume as the considered image
object. The proportions of the ellipsoid are equal to the proportions of the length, width
and thickness of the image object. The volume of the image object outside the ellipsoid is
compared with the volume inside the ellipsoid that is not lled out with the image object.
Parameters

v (x, y, z) is the elliptic distance at a pixel (x, y, z)

Pv is the set of pixels of an image object v
#Pv is the total number of pixels contained in Pv
Expression

= 2

# (x, y, z)Pv : v (x, y, z) 1
1
#Pv
Reference Book

11 June 2014

Object Features: Geometry

295

Figure 28.8. Elliptic fit of an image object v

Feature Value Range

[0, 1] ; 1 = complete tting, whereas 0 = < 50% voxels.

28.2.10 Main Direction [for 2D Image Objects]

Object Features > Geometry > Shape > Elliptic Fit > Main Direction
The Main Direction feature of an image object is dened as the direction of the eigenvector belonging to the larger of the two eigenvalues, derived from the covariance matrix of
the spatial distribution of the image object.
Parameters

VarX is the variance of X

VarY is the variance of Y
1 is the eigenvalue
Expression

180 1
tan (VarXY, 1 VarY ) + 90

Figure 28.9. The main direction is based on the direction of the larger eigenvector

Reference Book

11 June 2014

Object Features: Geometry

296

Feature Value Range

[0, 180]

28.2.11 Main Direction [for 3D Image Objects]

Object Features > Geometry > Shape > Main Direction
The Main Direction feature of a three-dimensional image object is computed as follows:
1. For each (2D) image object slice, the centers of gravities are calculated.
2. The co-ordinates of all centers of gravities a reused to calculate a line of best t,
according to the Weighted Least Square method.
3. The angle a between the resulting line of best t and the z-axis is returned as feature
value.

Figure 28.10. The line of best fit (blue) calculated from centers of gravity of image object

slices (light blue)

Feature Value Range

[0, 90]

28.2.12 Radius of Largest Enclosed Ellipse [for 2D Image Objects]

Object Features > Geometry > Shape > Radius of Largest Enclosed Ellipse

Reference Book

11 June 2014

Object Features: Geometry

297

The Radius of Largest Enclosed Ellipse feature describes how similar an image object is
to an ellipse. The calculation uses an ellipse with the same area as the object and based
on the covariance matrix. This ellipse is scaled down until it is totally enclosed by the
image object. The ratio of the radius of this largest enclosed ellipse to the radius of the
original ellipse is returned as feature value.
Parameters

v (x, y) is the elliptic distance at a pixel (x, y)

Expression

v (xo , yo ) , where (xo , yo ) = min v (x, y), (x, y)

/ Pv

Figure 28.11. Radius of the largest enclosed ellipse of in image object v , for a 2D image object

Feature Value Range

[0, ]

28.2.13 Radius of Largest Enclosed Ellipse [for 3D Image Objects]

Object Features > Geometry > Shape > Radius of Largest Enclosed Ellipse
The Radius of Largest Enclosed Ellipse feature describes how much the shape of an
image object is similar to an ellipsoid. The calculation is based on an ellipsoid with the
same volume as the object and based on the covariance matrix. This ellipsoid is scaled
down until it is totally enclosed by the image object. The ratio of the radius of this largest
enclosed ellipsoid to the radius of the original ellipsoid is returned as feature value.
Parameters

v (x, y, z) is the elliptic distance at a pixel (x, y, z)

Expression

v (xo , yo , zo ) , where (xo , yo , zo ) = min v (x, y, z), (x, y, z)

/ Pv
Reference Book

11 June 2014

Object Features: Geometry

298

Figure 28.12. Radius of the largest enclosed ellipse of in image object v , for a 3D image object

Feature Value Range

[0, ]

28.2.14 Radius of Smallest Enclosing Ellipse [for 2D Image Objects]

Object Features > Geometry >Shape > Radius of Smallest Enclosing Ellipse
The Radius of Smallest Enclosing Ellipse feature describes how much the shape of an
image object is similar to an ellipse. The calculation is based on an ellipse with the same
area as the image object and based on the covariance matrix. This ellipse is enlarged until
it encloses the image object in total. The ratio of the radius of this smallest enclosing
ellipse to the radius of the original ellipse is returned as feature value.

Figure 28.13. Radius of the smallest enclosed ellipse of in image object v , for a 2D image

object

Parameters

v (x, y) is the elliptic distance at a pixel (x, y)

Expression

v (xo , yo ) , where (xo , yo ) = max v (x, y), (x, y)

/ Pv

Reference Book

11 June 2014

Object Features: Geometry

299

Feature Value Range

[0, ]

28.2.15 Radius of Smallest Enclosing Ellipse [for 3D Image Objects]

Object Features > Geometry > Shape > Radius of Smallest Enclosing Ellipse
The Radius of Smallest Enclosing Ellipse feature describes how much the shape of an
image object is similar to an ellipsoid. The calculation is based on an ellipsoid with the
same volume as the image object and based on the covariance matrix. This ellipsoid is
enlarged until it encloses the image object in total. The ratio of the radius of this smallest
enclosing ellipsoid to the radius of the original ellipsoid is returned as feature value.

Figure 28.14. Radius of the smallest enclosed ellipse of in image object v , for a 3D image

object

Parameters

v (x, y, z) is the elliptic distance at a pixel (x, y, z)

Expression

v (xo , yo , zo ) , where (xo , yo , zo ) = max v (x, y, z), (x, y, z) Pv

Feature Value Range

[0, ]

28.2.16 Rectangular Fit [for 2D Image Objects]

Object Features > Geometry > Shape > Rectangular Fit
The Rectangular Fit feature describes how well an image object ts into a rectangle of
similar size and proportions. While 0 indicates no t, 1 indicates for a complete tting
image object.

Reference Book

11 June 2014

Object Features: Geometry

300

The calculation is based on a rectangle with the same area as the image object. The
proportions of the rectangle are equal to the proportions of the length to width of the
image object. The area of the image object outside the rectangle is compared with the
area inside the rectangle.
Parameters

v (x, y) is the elliptic distance at a pixel (x, y)

Expression

#(x, y)Pv : v (x, y) 1
#Pv

Figure 28.15. Rectangular fit of an image object v

Feature Value Range

[0, 1] ; where 1 is a perfect rectangle.

28.2.17 Rectangular Fit [for 3D Image Objects]

Object Features > Geometry > Shape > Rectangular Fit
The Rectangular Fit feature describes how well an image object ts into a cuboid of
similar size and proportions. While 0 indicates no t, 1 indicates a perfect t.
The calculation is based on a cuboid with the same volume as the considered image
object. The proportions of the cuboid are equal to the proportions of the length to width
to thickness of the image object. The volume of the image object outside the rectangle is
compared with the volume inside the cuboid that is not lled out with the image object.
Parameters

v (x, y, z) is the rectangular distance at a pixel (x, y, z)

#Pv is the total number of pixels contained in Pv

Reference Book

11 June 2014

Object Features: Geometry

301

Expression

#(x, y, z)Pv : v (x, y, z) 1
#Pv
Feature Value Range

[0, 1] ; 1 = complete tting, whereas 0 = 0% ts inside the rectangular approximation.

28.2.18 Roundness [for 2D Image Objects]

Object Features > Geometry > Shape > Roundness
The Roundness feature describes how similar an image object is to an ellipse. It is calculated by the difference of the enclosing ellipse and the enclosed ellipse. The radius of the
largest enclosed ellipse is subtracted from the radius of the smallest enclosing ellipse.
Parameters

is the radius of the smallest enclosing ellipse

max
v
is the radius of the largest enclosed ellipse
min
v
Expression

max
min
v
v

Figure 28.16. Roundness feature of image object v

Feature Value Range

[0, ] ; 0 = ideal.

Reference Book

11 June 2014

Object Features: Geometry

302

28.2.19 Roundness [for 3D Image Objects]

Object Features > Geometry > Shape > Roundness
The Roundness feature describes how much the shape of an image object is similar to an
ellipsoid. The more the shape of an image object is similar to an ellipsoid, the lower its
roundness.
It is calculated by the difference of the enclosing ellipsoid and the enclosed ellipsoid.
The radius of the largest enclosed ellipsoid is subtracted from the radius of the smallest
enclosing ellipsoid.
Parameters

is the radius of the smallest enclosing ellipsoid

max
v
is the radius of the largest enclosed ellipsoid
min
v
Expression

max
min
v
v
Feature Value Range

[0, ] ; 0 = ideal.

28.2.20 Shape Index [for 2D Image Objects]

Object Features > Geometry > Shape > Shape Index
The Shape index describes the smoothness of an image object border. The smoother the
border of an image object is, the lower its shape index. It is calculated from the Border
Length feature of the image object divided by four times the square root of its area.
Parameters

bvis the image object border length

4 #Pv is the border of square with area #Pv
Expression

b
v
4 #Pv
Feature Value Range

[1, ] ; 1 = ideal.

Reference Book

11 June 2014

Object Features: Geometry

303

Figure 28.17. Shape index of an image object v

28.2.21 Shape Index [for 3D Image Objects]

Object Features > Geometry > Shape > Shape Index
The Shape Index describes the smoothness of the surface of an image object. The
smoother the surface of an image object is, the lower its shape index. It is calculated
from the Border Length feature of the image object divided by the volume of the image
object.
Parameters

bv is the image object border length

Vv is the volume of image object v
Expression

bv
Vv
Feature Value Range

[0, 1]

28.3 To Super-Object
Object Features > Geometry > To Super-Object
Use the To Super-Object feature to describe an image object by its shape and relationship
to one of its superobjects, where appropriate. Editing the feature distance determines
which superobject is referred to. When working with thematic layers these features can
be of interest.

28.3.1 Rel. Area to Super-Object

Object Features > Geometry > To Super-Object > Rel. Area to Super-Object

Reference Book

11 June 2014

Object Features: Geometry

304

The area of an image object divided by the area of its superobject. If the feature value is
1, the image object is identical to its superobject. Use this feature to describe an image
object in terms of the amount of area it shares with its superobject.
Parameters

#Pv is the total number of pixels contained in Pv

#PUv(d) is the the size of the superobject of v in the image object level of the level
distance d
Expression

#Pv
#PUv(d)
Conditions

If Uv (d) = the formula is undened.

Feature Value Range

[0, 1]

28.3.2 Rel. Rad. Position to Super-Object

Object Features > Geometry > To Super-Object > Rel. Rad. Position to Super-Object
This value is calculated by dividing the distance between the center of a selected image
object and the center of its superobject, by the distance of the center of the most distant
image object (which has the same superobject). Use this feature to describe an image
object by its position relative to the center of its superobject.
Parameters

#Pv is the total number of pixels contained in Pv

#PUv(d) is the the size of the superobject of an image object v
dg (v,Uv(d)) is the distance of v to the center of gravity of the superobject Uv (d)
Expression

dg v,Uv (d)

max dg u,Uv (d)

Uv(d)(d)

Conditions

If Uv (d) = the formula is undened.

Reference Book

11 June 2014

Object Features: Geometry

305

Figure 28.18. Relative radial position to superobject

Feature Value Range

[0, 1]

28.3.3 Rel. Inner Border to Super-Object

Object Features > Geometry > To Super-Object > Rel. Inner Border to Super-Object
This feature is calculated by dividing the sum of the border shared with other image
objects, which have the same superobject, by the total border of the image object. If the
relative inner border to the superobject is 1, the image object of concern is not situated on
the border of its superobject. Use this feature to describe how much of an image object is
situated at the edge of its superobject.
Parameters

Nu (v) are neighbors of v that exist within the superobject Nu (v) : {u Nv : Uu (d)
Uv (d)}
bv is the image object border length
Expression

uNU (v) b(v, m)

bv
Conditions

If the feature range is 0 then v = Uv (d)

If the feature range is 1 then v is an inner object.
Feature Value Range

[0, 1]

Reference Book

11 June 2014

Object Features: Geometry

306

Figure 28.19. Relative inner border of an image object v to superobject u

28.3.4 Distance to Super-Object Center

Object Features > Geometry > To Super-Object > Distance to Super-Object Center
The distance of an image objects center to the center of its superobject. The calculated
distance is based upon segmentation on the image object level of the object of interest,
within the boundaries of the superobject.
Expression

dg (v,Uv (d)) is the distance of v to the center of gravity of the superobject Uv (d)
Feature Value Range

[0, sx sy]

Reference Book

11 June 2014

Object Features: Geometry

307

28.3.5 Elliptic Distance to Super-Object Center

Object Features > Geometry > To Super-Object > Elliptic Distance to Super-Object
Center
Distance of objects to the center of the superobject.
Expression

de (v,Uv (d))

Figure 28.20. Distance between the distance from the superobjects center to the center of a

sub-object

Feature Value Range

Typically [0, 5]

28.3.6 Is End of Super-Object

Object Features > Geometry > To Super-Object > Is End of Super-Object
This feature is true for two image objects a and b if following conditions are true:
a and b are sub-objects of the same superobject.
a is the image object with the maximum distance to the superobject.
b is the image object with the maximum distance to a .
Editable Parameter

Level Distance
Feature Value Range

[0, 1]
Reference Book

11 June 2014

Object Features: Geometry

308

28.3.7 Is Center of Super-Object

Object Features > Geometry > To Super-Object > Is Center of Super-Object
This feature is true if the image object is the center of its superobject.
Editable Parameter

Level Distance
Feature Value Range

[0, 1]

28.3.8 Rel. X Position to Super-Object

Object Features > Geometry > To Super-Object Rel. X Position to Super-Object
This feature returns the relative x position of an image object with regard to its superobject, based on the centers of gravity of both objects.
Editable Parameters

Level Distance the upward distance of image object levels in the image object
hierarchy between the image object and the superobject.
Expression

x = xCG of current image object xCG of superobject (where xCG is the center of gravity)
Feature Value Range

scene width scene width

+
2
2

28.3.9 Rel. Y Position to Super-Object

Object Features > Geometry > To Super-Object > Rel. Y Position to Super-Object
This feature returns the relative y position of an image object with regard to its superobject, based on the centers of gravity of both objects.
Editable Parameter

Level Distance: Upward distance of image object levels in the image object hierarchy between the image object and the superobject.

Reference Book

11 June 2014

Object Features: Geometry

309

Expression

y = yCG of current image object yCG of superobject (where yCG is the center of gravity)
Feature Value Range

scene height scene height

+
2
2

28.4 Based on Polygons

Object Features > Geometry > Based on Polygons
The polygon features provided by eCognition Developer 9.0 are based on the vectorization of the pixels that form an image object.

Figure 28.21. Raster image object (black area) with its polygon object (red lines) after vector-

ization

28.4.1 Edges Longer Than

Object Features > Geometry > Based on Polygons > Edges Longer Than
Editable Parameters

Minimum Length

28.4.2 Number of Right Angles With Edges Longer Than

Object Features > Geometry > Based on Polygons > Number of Right Angles with Edges
Longer Than
The number of right angles that have at least one side edge longer than a given threshold.
Editable Parameters

Minimum length

Reference Book

11 June 2014

Object Features: Geometry

310

Figure 28.22. A polygon with one rectangular angle

28.4.3 Area (Excluding Inner Polygons)

Object Features > Geometry > Based on Polygons > Area (Excluding Inner Polygons)
The Area (Excluding Inner Polygons) feature calculates the area of a polygon based on
Greens Theorem in a plane. In contrast to the Area (Including Inner Polygons) feature,
the feature value does not include the areas of any existing inner polygons.
Parameters

(xi , yi ), i = 0, . . . , n, with xo = xn and yo = yn as the given points

ai = xi yi+1 xi+1 yi
Expression

1 n1
ai
2 i=0

Figure 28.23. A polygon with an inner polygon that is not included in the feature value

Reference Book

11 June 2014

Object Features: Geometry

311

Feature Value Range

[0, scene size]

28.4.4 Area (Including Inner Polygons)

Object Features > Geometry > Based on Polygons > Area (Including Inner Polygons)
The Area (Excluding Inner Polygons) feature calculates the area of a polygon based on
Greens Theorem in a plane. Different to the Area (Excluding Inner Polygons) feature,
the feature value includes the areas of any existing inner polygons (for instance the single
polygon formed in the center of a donut-shaped object).

Figure 28.24. A polygon with an inner polygon that is included in the feature value

28.4.5 Average Length of Edges (Polygon)

Object Features > Geometry > Based on Polygons > average Length of Edges (Polygon)
The average length of all edges in a polygon.
Parameters

Xi is the length of edge i

n is the total number of edges
Expression
n

Average =

i=1

28.4.6 Compactness (Polygon)

Object Features > Geometry > Based on Polygons > Compactness (Polygon)
The ratio of the area of a polygon to the area of a circle with the same perimeter.

Reference Book

11 June 2014

Object Features: Geometry

312

Parameters

Area
Perimeter
Expression

4 Area
Perimeter2
Feature Value Range

[0, 1 for a circle]

28.4.7 Length of Longest Edge (Polygon)

Object Features > Geometry > Based on Polygons > Length of Longest Edge (Polygon)
The length of the longest edge of a polygon.

28.4.8 Number of Edges (Polygon)

Object Features > Geometry > Based on Polygons > Number of Edges (Polygon)
The number of edges of a polygon.

28.4.9 Number of Inner Objects (Polygon)

Object Features > Geometry > Based on Polygons > Number of Inner Objects (Polygon)
The number of inner polygons that are completely surrounded by a selected outer polygon.

28.4.10 Perimeter (Polygon)

Object Features > Geometry > Based on Polygons > Perimeter (Polygon)
The sum of the lengths of all the edges of a polygon.

28.4.11 Polygon Self-Intersection (Polygon)

Object Features > Geometry > Based on Polygons > Polygon Self-Intersection (Polygon)
The Polygon Self-Intersection (Polygon) feature allows the identication of a rare arrangement of image objects, leading to a polygon self-intersection when exported as a
polygon vector le.

Reference Book

11 June 2014

Object Features: Geometry

313

This feature enables you to identify the affected image objects and take measures To
avoid this self-intersection. All image objects with a value of 1 will cause a polygon
self-intersection when exported to a shapele.

Figure 28.25. This type of image object leads to a self-intersection at the circled point

To avoid the self-intersection, the enclosed image object needs to be merged with the
enclosing image object.
TIP: Use the Image Object Fusion algorithm to remove polygon intersections. To do so, set the domain to all image objects with a value larger
than 0 for the polygon intersection feature. In the algorithm parameter, set
the Fitting Function Threshold to Polygon Self-Intersection (Polygon) feature to zero and in the Weighted Sum group, set Target Value Factor to 1.
This will merge all image objects with a value of 1 for the Polygon SelfIntersection (Polygon) feature, so that the resulting image object will not
include a self-intersection.

Feature Value Range

[0, 1]

28.4.12 Std. Dev. of Length of Edges (Polygon)

Object Features > Geometry > Based on Polygons > Stddev of Length of Edges (Polygon)
The Std. Dev. of Length of Edges (Polygon) feature measures how the lengths of edges
deviate from their mean value.
Parameters

xi is the length of edge i

X is the mean value of all lengths
n is the total number of edges

Reference Book

11 June 2014

Object Features: Geometry

314

Expression

v
u n
u
u (xi x)
2
t i=1
n

28.5 Based on Skeletons

Object Features > Geometry > Based on Skeletons
For the better understanding of the following descriptions, the skeleton is structured in a
main line and subordinate branches. A node is a mid-point of the triangles created by the
Delaunay triangulation.

28.5.1 Number of Segments of Order

Object Features > Geometry > Based on Skeletons > Number of Segments of Order
The number of line segments of branches of an object that are of a given order. Note that
only segments that do not belong to a lower order are counted
Editable Parameter

Branch order: The main line of the skeleton has the order 0.
Feature Value Range

[0, depending on shape of objects]

28.5.2 Number of Branches of Order

Object Features > Geometry > Based on Skeletons > Number of Branches of Order
The number of branches of an object that are of a given order.
Editable Parameter

Branch order: The main line of the skeleton has the order 0
Feature Value Range

[0, depending on shape of objects]

Reference Book

11 June 2014

Object Features: Geometry

315

28.5.3 Average Length of Branches of Order

Object Features > Geometry > Based on Skeletons > average Length of Branches of
Order
The average length of branches of an object that are of a given order. The length of the
branch of the selected order is measured from the intersect point of the whole branch and
the main line to the end of the branch.
Editable Parameter

Branch order: The main line of the skeleton has the order 0.
Feature Value Range

[0, depending on shape of objects]

28.5.4 Number of Branches of Length

Object Features > Geometry > Based on Skeletons > Number of Branches of Length
The number of branches of an object that are of a special length up to a selected order.
All ends of branches are counted up to the selected order.
Editable Parameter

Branch order: The main line of the skeleton has the order 0.
Minimum length
Maximum length
Feature Value Range

[0; depending on shape of objects]

28.5.5 Average Branch Length

Object Features > Geometry > Based on Skeletons > Average Branch Length
The average length of all branches of an object.
Feature Value Range

[0, depending on shape of objects]

Reference Book

11 June 2014

Object Features: Geometry

316

28.5.6 Avrg. Area Represented by Segments

Object Features > Geometry > Based on Skeletons > Avrg. Area Represented by Segments
The average area of all triangles created by a Delaunay triangulation.
Feature Value Range

[0, depending on shape of objects]

28.5.7 Curvature/Length (Only Main Line)

Object Features > Geometry > Based on Skeletons > Curvature/Length (only main line)
The length-to-curvature ratio of the main line of an object. The curvature is the sum of
all changes in direction of the main line. Changes in direction are expressed by the acute
angle a in which sections of the main line, built by the connection between the nodes,
cross each other.

Figure 28.26. The main line (green) connects the mid-points of triagles (black and blue) cre-

ated by a Delaunay triangulation of the objects shape (not depicted)

Feature Value Range

[0, depending on shape of objects]

28.5.8 Degree of Skeleton Branching

Object Features > Geometry > Based on Skeletons > Degree of Skeleton Branching
The highest order of branching in an object.
Feature Value Range

[0, depending on shape of objects]

Reference Book

11 June 2014

Object Features: Geometry

317

28.5.9 Length of Main Line (No Cycles)

Object Features > Geometry > Based on Skeletons > Length of Main Line (No Cycles)
The sum of all distances between the nodes of the main line of an object. If an object
contains an island polygon a polygon derived from the inner borders of an image object
it is ignored and the main line may cross it (no cycles). This is different to the Length of
Main Line (regarding cycles) feature where the main line goes around the island polygon.
This feature does not visualize skeletons.
Feature Value Range

[0, depending on shape of objects]

28.5.10 Length of Main Line (Regarding Cycles)

Object Features > Geometry > Based on Skeletons > Length of Main Line (regarding
cycles)
The sum of all distances between the nodes of the main line of an object. If an object
contains an island polygon a polygon derived from the inner borders of an image object
the main line is calculated so as not to cross it (Regarding Cycles). In contrast to the
Length of Main Line (No Cycles) feature, the skeletons are visualized.
Feature Value Range

[0, depending on shape of objects]

28.5.11 Length/Width (Only Main Line)

Object Features > Geometry > Based on Skeletons >Length/Width (Only Main Line)
The length-to-width ratio of the main line of an object.
Feature Value Range

[0, depending on shape of objects]

28.5.12 Maximum Branch Length

Object Features > Geometry > Based on Skeletons > Maximum Branch Length
The length of the longest branch of an object. It is measured from the intersect point of
the branch and the main line to the end of the branch.

Reference Book

11 June 2014

Object Features: Geometry

318

Feature Value Range

[0, depending on shape of objects]

28.5.13 Number of Segments

Object Features > Geometry > Based on Skeletons > Number of Segments
The number of all segments of the main line and the branches of an object.
Feature Value Range

[0, depending on shape of objects]

28.5.14 Stddev Curvature (Only Main Line)

Object Features > Geometry > Based on Skeletons > Stddev Curvature (Only Main Line)
The standard deviation of the curvature is the result of the standard deviation of the
changes in direction of the main line. Changes in direction are expressed by the acute
angle in which sections of the mainline, built by the connection between the nodes, cross
each other.
Feature Value Range

[0, depending on shape of objects]

28.5.15 Stddev of Area Represented by Segments

Object Features > Geometry > Based on Skeletons > Stddev. of Area Represented by
Segments
The standard deviation of all triangles created by the Delaunay triangulation.
Feature Value Range

[0, depending on shape of objects]

28.5.16 Width (Only Main Line)

Object Features > Geometry > Based on Skeletons > Width (Only Main Line)
The width of an object based on the height of triangles created by a Delaunay triangulation. It is calculated by the average height h of all triangles crossed by the main line.
An exception is triangles where the height h does not cross one of the sides of the triangle.
In this case, the nearest side s is used to dene the height.

Reference Book

11 June 2014

Object Features: Geometry

319

Figure 28.27. Height h of an triangle that is crossed by the main line

Figure 28.28. Height of an triangle that is crossed by the main line. In this case the side s

defines the height

Feature Value Range

[0, depending on shape of objects]

Reference Book

11 June 2014

29 Object Features: Position

Object Features > Position
Position features refer to the position of an image object relative to the entire scene. These
features are of special interest when working with geographically referenced data, as an
image object can be described by its geographic position. Position features refer to the
pixel co-ordinate denition.

29.1 Distance
Object Features > Position > Distance

29.1.1 Distance to Line

Object Features > Position > Distance > Distance to Line
The distance between the center of gravity of a two-dimensional image object and a given
line. The line is dened manually by entering two points that are a part of this line. Note
that the line has neither a start nor an end.
Editable Parameters

To set, right-click the Distance to Line feature, select Edit Feature and adapt the coordinates of the two points:

First Coordinate (X)

First Coordinate (Y )
Second Coordinate (X)
Second Coordinate (Y )

Feature Value Range

[0, ]

320

Object Features: Position

321

Figure 29.1. Distance between an image object and a line

29.1.2 Distance to Scene Border

Object Features > Position > Distance > Distance to Scene Border
The distance of an image object slice (a 2D piece of the image object in a slice) to the
nearest border of the scene within the current slice. (The Z-direction is ignored.)
Parameters

min x is the minimum distance from the scene border at x -axis

max x is the maximum distance from the scene border at x -axis
min y is the minimum distance from the scene border at y -axis
max y is the maximum distance from the scene border at y -axis
(sx, sy) is the scene size

Expression

min{min x, sx max x, min y, sy max y}

Figure 29.2. Examples of the distance between an image object and the nearest scene border

Feature Value Range

[0, max sx 1, sy 1]

Reference Book

11 June 2014

Object Features: Position

322

29.1.3 T Distance to First Frame (Pxl)

Object Features > Position > Distance > T Distance to First Frame (Pxl)
The distance of an image object to the rst frame of the scene.

29.1.4 T Distance to Last Frame

Object Features > Position > Distance > T Distance to Last Frame
The distance of an image object to the last frame of the scene.

29.1.5 X Distance to Scene Left Border

Object Features > Position > Distance > X Distance to Scene Left Border
Horizontal distance of an image object slice (a 2D piece of the image object in a slice) to
the left border of the scene within the current slice.
Parameters

sx is the scene size at the x -axis

min x is the minimum distance from the scene border at x -axis
Expression

min x

(x,y)Pv

Figure 29.3. X-distance between the image object and the left border

Feature Value Range

[0, sx 1]

Reference Book

11 June 2014

Object Features: Position

323

29.1.6 X Distance to Scene Right Border

Object Features > Position > Distance > X Distance to Scene Right Border
Horizontal distance of an image object slice (a 2D piece of the image object in a slice) to
the right border of the scene within the current slice.
Parameters

sx is the scene size at the x -axis

max x is the maximum distance from the scene border at x -axis
Expression

sx max x
(x,y)Pv

Figure 29.4. X-distance to the image object right border

Feature Value Range

[0, sx 1]

29.1.7 Y Distance to Scene Bottom Border

Object Features > Position > Distance > Y Distance to Scene Bottom Border
Vertical distance of an image object slice (a 2D piece of the image object in a slice) to the
bottom border of the scene within the current slice.
Parameters

sy is the scene size

min y is the minimum distance from the scene border at y -axis
Expression

min y

(x,y)Pv

Reference Book

11 June 2014

Object Features: Position

324

Figure 29.5. Y-distance between the image object and the bottom border

Feature Value Range

[0, sy 1]

29.1.8 Y Distance to Scene Top Border

Object Features > Position > Distance > Y Distance to Scene Top Border
Vertical distance of an image object slice (a 2D piece of the image object in a slice) to the
top border of the scene within the current slice.
Parameters

sy is the scene size

max y is the maximum distance from the scene border at y -axis
Expression

sy max y
(x,y)Pv

Figure 29.6. Y-distance between the image object and the top border of the scene

Feature Value Range

[0, sy 1]
Reference Book

11 June 2014

Object Features: Position

325

29.1.9 Z Distance to First Slice (Pxl)

Object Features > Position > Distance > Z Distance to First Slice (Pxl)
Distance of an image object to the rst slice of the scene.

29.1.10 Z Distance to Last Slice (Pxl)

Object Features > Position > Distance > Z Distance to Last Slice (Pxl)
Distance of an image object to the last slice of the scene.

29.2 Coordinate
Object Features > Position > Coordinate

29.2.1 Is at Active Pixel

Object Features > Position > Coordinate > Is at Active Pixel
This feature returns 1 if the current active pixel is within the image object, otherwise 0.

29.2.2 Time (Pxl)

Object Features > Position > Coordinate > Time
T-position of the center of an image object. The calculation is based on the center of
gravity (geometric center) of the image object in the internal map.
Parameters

tv is the t -center of an image object v

xv is the x -center of an image object v in an internal map with xv (map) =
(xmap ,ymap )Pv xmap
sxframe is the extent in x of each slice and frame

1
#Pv

Expression

v (map)
tv = oor xsx
frame
Feature Value Range

[0.5, number of frames in map 0.5]

Reference Book

11 June 2014

Object Features: Position

326

29.2.3 Time Max (Pxl)

Object Features > Position > Coordinate > Time Max (Pxl)
Maximum t -position of an image object derived from its bounding box. The calculation
is based on the maximum t -position of the image object in the internal map.
Parameters

tmax (v) is the maximum t -position of an image object v

xmax (v, map) is the maximum x -position of an image object v in an internal map
sx f rame is the extent in x of each slice and frame
Expression

xv (map)
tv = oor sxframe
Feature Value Range

[1, number of frames in map]

29.2.4 Time Min (Pxl)

Object Features > Position > Coordinate > Time Min (Pxl)
Minimum t -position of an image object derived from its bounding box. The calculation
is based on the minimum t -position of the image object in the internal map.
Parameters

tmin (v) is the minimum t -position of an image object v

xmin (v, map) is the minimum x -position of an image object t in an internal map
sxframe is the extent in x of each slice and frame
Expression

xmin (v,map)
tmin (v) = oor
sxframe
Feature Value Range

[0.5, number of frames in map 1]

Reference Book

11 June 2014

Object Features: Position

327

29.2.5 X Center
Object Features > Position > Coordinate > X Center
X-position of the center of an image object. The calculation is based on the center of
gravity (geometric center) of the image object in the internal map.
Parameters

xv is the x -center of an image object v

xv (map) is the x -center of an image object v in an internal map with xv (map) =
1
#Pv (xmap ,ymap )Pv xmap
#Pv is the total number of pixels contained in Pv
(xmap , ymap ) are the coordinates in an internal map
sxframe is the extent in x of each slice and frame
Expression

xv = xv (map) oor

xv (map)
sxframe

sxframe

Feature Value Range

[scene extent in x, 0.5]

29.2.6 X Max.
Object Features > Position > Coordinate > X Max
Maximum x -position of an image object derived from its bounding box. The calculation
is based on the maximum x -position of the image object in the internal map.
Parameters

xmax (v) is the minimum x -position of an image object v

xmax (v, map) is the maximum x -position of an image object t in an internal map
sxframe is the extent in x of each slice and frame
Expression

xmax (map)
xmax (v) = xmax (v, map) oor sxframe
sxframe
Feature Value Range

[1, scene extent in x]

Reference Book

11 June 2014

Object Features: Position

328

Figure 29.7. Maximum value of the x-coordinate at the image object border

29.2.7 X Min.
Object Features > Position > Coordinate > X Min
Minimum x-position of an image object derived from its bounding box. The calculation
is based on the minimum x-position of the image object in the internal map.
Parameters

xmin (v) is the minimum x -position of an image object v

xmin (v, map) is the minimum x -position of an image object v in an internal map
sxframe is the extent in x of each slice and frame
Expression

xmin (v,map)
sxframe
xmin (v) = xmin (v, map) oor
sxframe

Figure 29.8. Minimum value of the x-coordinate at the image object border

Feature Value Range

[0, scene extent in x 1]

29.2.8 Y Center
Object Features > Position > Coordinate > Y Center
Y-position of the center of an image object. The calculation is based on the center of
gravity (geometric center) of the image object in the internal map.

Reference Book

11 June 2014

Object Features: Position

329

Parameters

yv is the y -center of an image object v

yv (map) is the y -center of an image object v in an internal map with yv (map) =
1
#Pv (xmap ,ymap )Pv ymap
#Pv is the total number of pixels contained in Pv
(xmap , ymap) are the coordinates in an internal map
syslice is the extent in y of each slice and frame
Expression

yv (map)
yv = yv (map) oor syslice syslice

Figure 29.9. Center of gravity of an image object

Feature Value Range

[0.5, scene extent in x 0.5]

29.2.9 Y Max.
Object Features > Position > Coordinate > Y Max
Maximum y-position of an image object derived from its bounding box. The calculation
is based on the maximum y-position of the image object in the internal map.
Parameters

ymax (v) is the maximum y -position of an image object v

ymax (v, map) is the minimum y -position of an image object v in an internal map
syslice is the extent in y of each slice and frame
Expression

ymax (v) = ymax (v, map) oor

ymax (v,map)
syslice

syslice

Reference Book

11 June 2014

Object Features: Position

330

Figure 29.10. Maximum value of the y-coordinate at the image object border

Feature Value Range

[1, scene extent in y]

29.2.10 Y Min.
Object Features > Position > Coordinate > Y Min
Minimum y-position of an image object derived from its bounding box. The calculation
is based on the minimum y-position of the image object in the internal map.
Parameters

ymin (v) is the minimum y -position of an image object v

ymin (v, map) is the minimum y -position of an image object v in an internal map
syslice is the extent in y of each slice and frame
Expression

syslice
ymin (v) = ymin (v, map) oor yminsy(v,map)
slice

Figure 29.11. Minimum value of the y-coordinate at the image object border

Feature Value Range

[0, scene extent in y 1]

Reference Book

11 June 2014

Object Features: Position

331

29.2.11 Z Center
Object Features > Position > Coordinate > Z Center
Z -position of the center of an image object. The calculation is based on the center of
gravity (geometric center) of the image object in the internal map.
Parameters

zv is the z -center of an image object v

yv (map) is the y -center of an image object v in an internal map with yv (map) =
1
#Pv (xmap ,ymap )Pv ymap
syslice is the extent in y of each slice and frame
Expression

yv (map)
zv = oor syslice
Feature Value Range

[0.5, number of slices in map 0.5]

29.2.12 Z Max
Object Features > Position > Coordinate > Z Max
Maximum z -position of an image object derived from its bounding box. The calculation
is based on the maximum z -position of the image object in the internal map.
Parameters

zmax (v) is the maximum z -position of an image object v

ymax (v, map) is the minimum y -position of an image object v in an internal map
syslice is the extent in y of each slice and frame
Expression

zmax (v) = oor

ymax (v,map)
syslice

Feature Value Range

[1, number of slices in map]

Reference Book

11 June 2014

Object Features: Position

332

29.2.13 Z Min
Object Features > Position > Coordinate > Z Min
Minimum z -position of an image object derived from its bounding box. The calculation
is based on the minimum z -position of the image object in the internal map.
Parameters

zmin (v) is the maximum z -position of an image object v

ymin (v, map) is the minimum y -position of an image object v in an internal map
syslice is the extent in y of each slice and frame
Expression

ymin (v,map)
zmin (v) = oor
syslice
Feature Value Range

[0, number of slices in map 1]

29.3 Is Object in Region

Object Features > Position > Is Object in Region
The Is Object In Region feature checks if an image object is located in a given region. If
this is true, the feature value is 1 (= true), otherwise it is 0 (= false).

29.3.1 Editable Parameters

Region

29.3.2 Feature Value Range

[0, 1]

Reference Book

11 June 2014

30 Object Features: Texture

Object Features > Texture
Texture features are used to evaluate the texture of image objects and include features
based on an analysis of sub-objects helpful for evaluating highly textured data. In addition, features based upon the gray level co-occurrence matrix after Haralick are available.
Features are structured in the following groups:
Texture concerning the spectral information of the sub-objects
Texture concerning the form of the sub-objects
Texture after Haralick based on the gray level co-occurrence matrix (GLCM),
which is a tabulation of how often different combinations of pixel gray levels occur
in an image.

30.1 Layer Value Texture Based on Sub-objects

Object Features > Texture > Layer Value Texture Based on Sub-objects
Layer Value Texture Based on Sub-objects features refer to the spectral information provided by the image layers. To refer to an image object level of sub-objects, you can edit
the level distance.

30.1.1 Mean of Sub-objects: Std. Dev.

Object Features > Texture > Layer Value Texture Based on Sub-objects > Mean of Subobjects: Stddev.
Standard deviation of the different layer mean values of the sub-objects. This feature
might appear very similar to the simple standard deviation computed from the single pixel
values (layer values); however it can be more meaningful because assuming an adequate
segmentation the standard deviation is computed over homogeneous and meaningful
areas.
The smaller the sub-objects, the more the feature value approaches the standard deviation
calculated from single pixels.
Parameters

Sv (d) is the sub-object of an image object v at distance d

333

Object Features: Texture

334

ck (u) is the mean intensity of layer k of an image object u

d is the image object level distance
Expression

v
u
u
t

1
#Sv (d)

(ck

uSv (d)

(u))2

!
1
ck (u) ck (u)
#Sv (d) uS(d)
uS (d)
v

Feature Value Range

[0, depending on bit depth of data]

Conditions

If Sv (d) = the expression is invalid.

30.1.2 Avrg. Mean Diff. to Neighbors of Sub-objects

Object Features > Texture > Layer Value Texture Based on Sub-objects > Avrg. Mean
Diff. to Neighbors of Sub-objects
The contrast inside an image object expressed by the average mean difference of all its
sub-objects for a specic layer.
This feature has a certain spatial reference, because a local contrast inside the area covered by the image object is described. For each single sub-object the layer L mean difference (absolute values) to adjacent sub-objects of the same superobject is calculated. The
feature value is the mean value of the layer L mean differences.
Parameters

Sv (d) is the sub-object of an image object v at distance d

k (u) is the mean difference to neighbor of layer k of an image object u
d is the image obejct level distance

Expression

1
k (u)

#Sv (d) uSv (d)

Feature Value Range

[0, depending on bit depth of data]

Reference Book

11 June 2014

Object Features: Texture

335

Conditions

If Sv (d) = the formula is invalid

30.2 Shape Texture Based on Sub-objects

Object Features > Texture > Shape Texture Based on Sub-objects
Shape Texture Based on Sub-objects features refer to the shape of sub-objects. To use
these features successfully we recommend an accurate segmentation of the map to ensure
sub-objects are as meaningful as possible. To refer to an image object level of sub-objects,
you can edit the level distance.

30.2.1 Area of Sub-objects: Mean

Object Features > Texture > Shape Texture Based on Sub-objects > Area of Sub-objects:
Mean
Mean value of the areas of the sub-objects.
Parameters

Sv (d) is the sub-object of an image object v at distance d

#Pu is the total number of pixels contained in u
d is the image object level distance
Expression

1
#Pu
#Sv (d) S(d)
v

Feature Value Range

[0, scene size]

Condition

If Sv (d) = the formula is invalid.

30.2.2 Area of Sub-objects: Std. Dev.

Object Features > Texture > Shape Texture Based on Sub-objects > Area of Sub-objects:
Std Dev.
Standard deviation of the areas of the sub-objects.

Reference Book

11 June 2014

Object Features: Texture

336

Parameters

Sv (d) is the sub-object of an image object v at distance d

Pu is the total number of pixels contained in u
d is the image object level distance
Expression

v
u
u
t

1
#Sv (d)

1
#Pu2 #Sv (d) #Pu #Pu
uS (d)
uS (d)
uS (d)
v

Feature Value Range

[0, scene size]

Condition

If Sv (d) = the formula is invalid.

30.2.3 Density of Sub-objects: Mean

Object Features > Texture > Shape Texture Based on Sub-objects > Density of Subobjects: Mean
Mean value calculated from the densities of the sub-objects.
Parameters

Sv (d) is the sub-object of an image object v at distance d

a(u) is the density of u
d is the image object level distance
Expression

1
a(u)
#Sv (d) uS(d)
v

Feature Value Range

[0, depending on image object shape]

Condition

If Sv (d) = the formula is invalid.

Reference Book

11 June 2014

Object Features: Texture

337

30.2.4 Density of Sub-objects: Std. Dev.

Object Features > Texture > Shape Texture Based on Sub-objects > Density of Subobjects: Stddev.
Standard deviation calculated from the densities of the sub-objects.
Parameters

Sv (d) is the sub-object of an image object v at distance d

a(u) is the density of u
d is the image object level distance
Expression

v
u
u
t

!
1
a2 (u) #Sv (d) a(u) a(u)
uS (d)
uS (d)
uS (d)

1
#Sv (d)

Feature Value Range

[0, depending on image object shape]

Condition

If Sv (d) = the formula is invalid.

30.2.5 Asymmetry of Sub-objects: Mean

Object Features > Texture > Shape Texture Based on Sub-objects > Asymmetry of Subobjects: Mean
Mean value of the asymmetries of the sub-objects.
Parameters

Sv (d) is the sub-object of an image object v at distance d

a(u) is the asymmetry of u
d is the image object level distance
Expression

1
a(u)
#Sv (d) uS(d)
v

Reference Book

11 June 2014

Object Features: Texture

338

Feature Value Range

[0, depending on image object shape]

Condition

If Sv (d) = the formula is invalid.

30.2.6 Asymmetry of Sub-objects: Std. Dev.

Object Features > Texture > Shape Texture Based on Sub-objects > Asymmetry of Subobjects: Stddev.
Standard deviation of the asymmetries of the sub-objects.
Parameters

Sv (d) is the sub-object of an image object v at distance d

a(u) is the asymmetry of u
d is the image object level distance
Expression

1
a(u)
#Sv (d) uS(d)
v

Feature Value Range

[0, depending on image object shape]

Condition

If Sv (d) = the formula is invalid.

30.2.7 Direction of Sub-objects: Mean

Object Features > Texture > Shape Texture Based on Sub-objects > Direction of Subobjects: Mean
The mean value of the directions of the sub-objects. In the computation, the directions
are weighted with the asymmetry of the respective sub-objects. The more asymmetric an
image object, the more signicant its main direction.
Before computing the actual feature value, the algorithm compares the variance of all
sub-object main directions with the variance of the sub-object main directions, where
all directions between 90 and 180 are inverted. The set of sub-object main directions
which has the lower variance is selected for the calculation of the main direction mean
value, weighted by the sub-object asymmetries.
Reference Book

11 June 2014

Object Features: Texture

339

Figure 30.1. For calculation, the directions between 90 and 180 are inverted, which means

the direction is reduced by 180

Parameters

Sv (d) is the sub-object of an image object v at distance d

a(u) is the main direction of u
d is the image object level distance
Expression

1
a(u)
#Sv (d) uS(d)
v

Feature Value Range

[0 , 180 ]
Condition

If Sv (d) = the formula is invalid.

30.2.8 Direction of Sub-objects: Std. Dev.

Object Features > Texture > Shape Texture Based on Sub-objects > Direction of Subobjects: Stddev
Standard deviation of the directions of the sub-objects. Similar to Direction of Subobjects: Mean feature, the sub-object main directions are weighted by the asymmetries of
the respective sub-objects. The set of sub-object main directions from which the standard
deviation is calculated is determined in the same way.

30.3 Texture After Haralick

Object Features > Texture > Texture After Haralick
Reference Book

11 June 2014

Object Features: Texture

340

The gray level co-occurrence matrix (GLCM) is a tabulation of how often different combinations of pixel gray levels occur in a scene. A different co-occurrence matrix exists for
each spatial relationship. To receive directional invariance, the sum of all four directions
(0, 45, 90, 135) are calculated before texture calculation. An angle of 0 represents
the vertical direction, an angle of 90 the horizontal direction. In Trimble software, texture after Haralick is calculated for all pixels of an image object. To reduce border effects,
pixels bordering the image object directly (surrounding pixels with a distance of 1) are
additionally taken into account.

The normalized GLCM is symmetrical. The diagonal elements represent pixel pairs with
no gray level difference. Cells, which are one cell away from the diagonal, represent pixel
pairs with a difference of only one gray level. Similarly, values in cells that are two pixels
away from the diagonal, show how many pixels have two gray levels and so forth. The
more distant to the diagonal, the greater the difference between the pixels gray levels is.
Summing-up the values of these parallel diagonals, gives the probability for each pixel to
be 0, 1, 2 or 3 etc. different from its neighbor pixels.

Another approach to measure texture is to use a gray-level difference vector (GLDV)

instead of the GLCM. The GLDV is the sum of the diagonals of the GLCM. It counts the
occurrence of references to the neighbor pixels absolute differences. In Trimble software
the GLCM and GLDV are calculated based on the pixels of an object. They are computed
for each input layer. Within each Texture after Haralick feature you have the choice of
either one of the above directions or of all directions.
The calculation of Texture after Haralick is independent of the image datas bit-depth.
The dynamic range is interpolated to 8 bit before evaluating the co-occurrence. However,
Reference Book

11 June 2014

Object Features: Texture

341

if 8 bit data is used directly the results will be most reliable. When using data of higher
dynamic than 8 bit, the mean and standard deviation of the values is calculated. Assuming
a Gaussian distribution of the values, a value greater than 95% is inbetween the interval,
which is subdivided into 255 equal sub-intervals to obtain an 8-bit representation.
x 3 < x < x + 3
The result of any texture after Haralick analysis is dependent upon the direction of the
analysis, whether it is All Directions, Direction 0, Direction 45, Direction 90 or Direction 135. In addition, each feature is calculated based upon the gray values of one
selectable layer.
The calculation of any Haralick texture feature is very processor intensive because of the
calculation of the GLCM; as a result, performance may suffer.

30.3.1 Calculation of GLCM

Image object pixels are scanned and put into a 256 x 256 matrix, with pixel value and
neighboring pixel value as co-ordinates. The value of the matrix is a normalized number
of pair occurrences (number of occurrences number of all pairs). So GLCM matrix
values are always within the range [0, 1] . The pixel value is either the pixel value of a
particular layer or the mean value of all layers (depending on the Haralick feature chosen).
The neighboring pixel calculation depends upon the feature angle:

0 (x; y 1) and (x; y + 1)

45 (x 1, y 1) and (x + 1, y + 1)
90 (x 1; y) and (x + 1; y)
135 (x 1, y + 1) and (x + 1, y 1)
All dir all neighbor pixels described above.

The actual Haralick feature is calculated using the GLCM matrix and algorithm of a
particular feature. For example, the contrast feature is calculated using the formulae:
for (int i=0; i<mtrx.Rows(); i++)
for (int j=0; j<mtrx.Cols(); j++)
ctrst += mtrx[i][j] * ((i-j)*(i-j)).
TIP: For each Haralick texture feature there is a performance optimized
version labeled quick 8/11. The performance optimization works only on
data with a bit depth of 8 bit or 11-bit. Hence the label quick 8/11. Use the
performance optimized version whenever you work with 8- or 11-bit data.
For 16 bit data, use the conventional Haralick feature.

30.3.2 Parameters

i is the row number

j is the column number
Vi, j is the value in the cell i, j of the matrix
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns
Reference Book

11 June 2014

Object Features: Texture

342

30.3.3 Expression
Every GLCM is normalized according to the following operation:
Pi, j =

Vi, j
N1

Vi, j

i, j=0

30.3.4 References
Haralick RM, Shanmugan K and Dinstein I (1973). Textural Features for Image Classication. IEEE Transactions on Systems, Man and Cybernetics. Vol SMC-3, No. 6 pp.
610621
Haralick RM (1979). Statistical and Structural Approaches to Texture. Proceedings of
the IEEE. Vol. 67, No. 5, pp. 786804
Conner RW and Harlow CA (1980). A Theoretical Comparison of Texture Algorithms.
IEEE Transactions on Pattern Analysis and Machine Intelligence. Vol PAMI-2, No. 3.

30.3.5 GLCM Homogeneity

Object Features > Texture > Texture After Haralick > GLCM Homogeneity
If the image is locally homogeneous, the value is high if GLCM concentrates along the
diagonal. Homogeneity weights the values by the inverse of the contrast weight with
weights, decreasing exponentially according to their distance to the diagonal.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

Pi, j
2
i, j=0 1 + (i j)

Feature Value Range

[0, 1]

Reference Book

11 June 2014

Object Features: Texture

343

30.3.6 GLCM Contrast

Object Features > Texture > Texture After Haralick > GLCM Contrast
Contrast is the opposite of homogeneity. It is a measure of the amount of local variation
in the image. It increases exponentially as i j increases.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

i, j=0

Pi, j (i j)2

Feature Value Range

[0, 65025]

30.3.7 GLCM Dissimilarity

Object Features > Texture > Texture After Haralick > GLCM dissimilarity
Similar to contrast, but increases linearly. High if the local region has a high contrast.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

i, j=0

Pi, j |i j|

Feature Value Range

[0, 255]

Reference Book

11 June 2014

Object Features: Texture

344

30.3.8 GLCM Entropy

Object Features > Texture > Texture After Haralick > GLCM Entropy
The value for entropy is high, if the elements of GLCM are distributed equally. It is low
if the elements are close to either 0 or 1. Since ln(0) is undened, it is assumed that
0 ln(0) = 0.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

Pi, j ( ln Pi, j )

i, j=0

Feature Value Range

[0, 10404]

30.3.9 GLCM Ang. 2nd Moment

Object Features > Texture > Texture After Haralick > GLCM Ang. 2nd Moment
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

(Pi, j )2

i, j=0

Feature Value Range

[0, 1]

Reference Book

11 June 2014

Object Features: Texture

345

30.3.10 GLCM Mean

Object Features > Texture > Texture After Haralick > GLCM Mean
The GLCM mean is the average expressed in terms of the GLCM. The pixel value is not
weighted by its frequency of occurrence itself, but by the frequency of its occurrence in
combination with a certain neighbor pixel value.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns

Expression
N1

i, j =

Pi, j

i, j=0

Feature Value Range

[0, 255]

30.3.11 GLCM Std. Dev.

Object Features > Texture > Texture After Haralick > GLCM Std Dev
GLCM standard deviation uses the GLCM, therefore it deals specically with the combinations of reference and neighbor pixels. Thus, it is not the same as the simple standard
deviation of gray levels in the original image.
Calculating the standard deviation using i or j gives the same result, since the GLCM is
symmetrical.
Standard deviation is a measure of the dispersion of values around the mean. It is similar
to contrast or dissimilarity.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns
i, j is the GLCM mean

Reference Book

11 June 2014

Object Features: Texture

346

Expression

2i, j =

i, j=0

Pi, j (i, j i, j )

Standard Deviation

2i, j

Feature Value Range

[0, 255]

30.3.12 GLCM Correlation

Object Features > Texture > Texture After Haralick > GLCM Correlation
Measures the linear dependency of gray levels of neighboring pixels.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns
i, j is the GLCM mean
i, j is the GLCM standard deviation

Expression
N1

(i, i )( j j )
Pi, j q
i, j=0
(2i )(2i )

Feature Value Range

[0, 1]

30.3.13 GLDV Angular 2nd Moment

Object Features > Texture > Texture After Haralick > GLDV Ang. 2nd Moment
Similar to GLCM Angular Second Moment: it measures the local homogeneity. The
value is high if some elements are large and the remaining ones are small.

Reference Book

11 June 2014

Object Features: Texture

347

Parameters

N is the number of rows or columns

Vk is the image object level, k = 1, . . . n
Expression
N1

Vk2

k=0

Feature Value Range

[0, 1]

30.3.14 GLDV Entropy

Object Features > Texture > Texture After Haralick > GLDV Entropy
The values are high if all elements have similar values. It is the opposite of GLDV
Angular Second Moment.
Parameters

i is the row number

j is the column number
Pi, j is the normalized value in the cell i, j
N is the number of rows or columns
Vk is the image object level, k = 1, . . . n

Expression

Since ln(0) is undened, it is assumed that 0 ln(0) = 0.

Vk (lnVk )
k=0

Feature Value Range

[0, 10404]

30.3.15 GLDV Mean

Object Features > Texture > Texture After Haralick > GLDV Mean
The mean is mathematically equivalent to GLCM Dissimilarity on page 343. It is only
left here for compatibility reasons.

Reference Book

11 June 2014

Object Features: Texture

348

Parameters

N is the number of rows or columns

Vk is the image object level, k = 1, . . . n
Expression
N1

k(Vk )
k=0

Feature Value Range

[0, 255]

30.3.16 GLDV Contrast

Object Features > Texture > Texture After Haralick > GLDV Contrast
It is mathematically equivalent to the GLCM Contrast measure, described previously. It
is still included for compatibility reasons.
Parameters

N is the number of rows or columns

Vk is the image object level, k = 1, . . . n
Expression
N1

Vk k2

k=0

Feature Value Range

[0, 65025]

30.3.17 GLCM/GLDV . . . (Quick 8/11)

Object Features > Texture > Texture After Haralick GLCM/GLDV [quick 8/11]
For each Haralick texture feature there is a performance optimized version with a quick
8/11 sufx.
They are labeled quick 8/11 because the performance optimization works only on data
with a bit depth of 8 or 11-bits. Use the performance optimized version whenever you
work with 8- or 11-bit data; for 16 bit data, use the conventional Haralick feature.

Reference Book

11 June 2014

31 Object Features: Variables

Object Features > Variables
Object Variables are local variables. In contrast to scene variables they store values for
each individual image object. Think of it as each image object having its own version of
the variable. There is one instance per image object in a project.

31.1 Object Variable

Object Features > Variables > Object Variable
If existing, all object variables are listed in the feature tree.
You can dene a new object variable by clicking on Create new Object Variable to store
interim values for each individual image object.

31.1.1 Editable Parameters

Name
Value: Insert an initial value for the variable.
Choose whether the new variable is numeric (Double) or textual (String)
Shared: Select if you want to share the new variable among different rule sets

349

32 Object Features: Hierarchy

Object Features > Hierarchy
Hierarchy features provide information about the embedding of an image object within
the image object hierarchy.

32.1 Level
Object Features > Hierarchy > Level
Returns the level object associated with an image object.

32.2 Level Number

Object Features > Hierarchy > Level Number
The number of the image object level an image object is situated in.
This feature is helpful when performing classications on different image object levels to
dene which class description is valid for which image object level.

32.2.1 Parameters
Uv (d) are superobjects of an image object v at distance d

32.2.2 Expression
minUv (d) = 0
d

32.2.3 Feature Value Range

[1, number of image object levels]

32.2.4 Conditions
This feature requires more than one image object level.
350

Object Features: Hierarchy

351

32.3 Number of Higher Levels

Object Features > Hierarchy > Number of Higher Levels

32.3.1 Parameters
d is the distance between neighbors
Sv (d) are sub-objects of an image object v at a distance d

32.3.2 Expression

minSv (d) = 1
d

32.3.3 Feature Value Range

[1, number of image object levels 1]

32.4 Number of Neighbors

Object Features > Hierarchy > Number of Neighbors
The number of the direct neighbors of an image object (in other words, neighbors with
which it has a common border) on the same image object level in the image object hierarchy.

32.4.1 Parameters
Nv (d) are neighbors of an image object v at a distance d

32.4.2 Expression
#Nv (d)

32.4.3 Feature Value Range

[0, number of pixels of entire scene]

32.5 Number of Sub-Objects

Object Features > Hierarchy > Number of Sub-objects
Concerning an image object, the number of sub-objects that are located on the next lower
image object level in the image object hierarchy.

Reference Book

11 June 2014

Object Features: Hierarchy

352

32.5.1 Parameters
Sv (d) are sub-objects of an image object v at a distance d

32.5.2 Expression
#Sv (d)

32.5.3 Feature Value Range

[0, number of pixels of entire scene]

32.6 Number of Sublevels

Object Features > Hierarchy > Number of Sublevels
The number of image object levels situated below a given image object.

32.6.1 Parameters
d is the distance between neighbors
Uv (d) are superobjects of an image object v at a distance d

32.6.2 Expression

minUv (d) = 1
d

32.6.3 Feature Value Range

[1, number of image object levels 1]

Reference Book

11 June 2014

33 Object Features: Thematic

Attributes
Object Features > Thematic Attributes
Thematic attribute features are used to describe an image object using information provided by thematic layers. If your scene contains a thematic layer, its thematic objects
properties, can be used to create thematic attribute features, which can be used for developing ruleware. Depending on the attributes of the thematic layer, a large range of
different features becomes available.

33.1 Minimum overlap [%] with thematic polygons

Object Features > Thematic Attributes > Minimum overlap [%] with thematic polygons
The feature computes the minimum value of the overlap between an image object and a
selected vector layer in percent. If the image object overlaps with more than one vector
object of the selected layer the feature returns only its minimum overlap value.

33.1.1 Editable Parameter

Thematic Layer

33.1.2 Feature Value Range

[0, 100]

33.2 Maximum overlap [%] with thematic polygons

Object Features > Thematic Attributes > Maximum overlap [%] with thematic polygons
The feature computes the maximum value of the overlap between an image object and a
selected vector layer in percent. If the image object overlaps with more than one vector
object of the selected layer the feature returns only its maximum overlap value.

353

Object Features: Thematic Attributes

354

33.2.1 Editable Parameter

Thematic Layer

33.2.2 Feature Value Range

[0, 100]

33.3 Number of Overlapping Thematic Objects

Object Features > Thematic Attributes > Number of Overlapping Thematic Objects
The number of thematic objects that an image object overlaps with. The scene contains a
thematic layer, otherwise the feature value is 0.

33.3.1 Editable Parameter

Thematic Layer

33.3.2 Feature Value Range

[0, number of thematic objects]

33.4 Thematic Object Attribute

Object Features > Thematic Attributes > [Thematic Object Attribute]
If existing, Thematic Objects Attribute features referring to a thematic layer are listed in
the feature tree. Preconditions:
The scene includes a thematic layer, otherwise the feature value is 0.
Available only for image objects that overlap with one thematic object, otherwise
the feature value is undened.

33.4.1 Editable Parameters

Thematic Layer
Thematic Layer Attribute

Reference Book

11 June 2014

34 Object Features: Point Cloud

Features
Object Features > Point Cloud Features
Features in this group are descriptive structure statistics, calculated from point clouds.
For more information on the parameters, see LiDAR Algorithms (p 163).

34.1 Intensity
The point intensity values.

34.1.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last

355

Object Features: Point Cloud Features

356

You can dene to ignore elevation below the specied value.

Enter the value manually or choose:
Minimum allowed elevation

<none>
From Feature
From Array
You can dene to ignore elevation above the specied value.
Enter the value manually or choose:
Maximum allowed elevation

<none>
From Feature
From Array
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
Reference Book

11 June 2014

Object Features: Point Cloud Features

357

From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.2 X coordinate
The X coordinate value of each point.

34.2.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

358

Point Filter - By RGB value

Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of blue value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.3 Y coordinate
The Y coordinate value of each point.

34.3.1 Parameters
Image Layer (LAS File Source)

Reference Book

11 June 2014

Object Features: Point Cloud Features

359

Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of blue value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

360

Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.4 Elevation
The point elevation values.

34.4.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
Reference Book

11 June 2014

Object Features: Point Cloud Features

361

All not selected classes

Point Filter - By RGB value
Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of blue value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.5 Return Number

The Return Number is the pulse return number for a given output pulse.

34.5.1 Parameters
Image Layer (LAS File Source)

Reference Book

11 June 2014

Object Features: Point Cloud Features

362

Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

363

Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.6 Number of Returns

The Number of Returns is the total number of returns for a given pulse.

34.6.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
Reference Book

11 June 2014

Object Features: Point Cloud Features

364

All not selected classes

Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.7 Red
The Red image channel value associated with this point.

34.7.1 Parameters
Image Layer (LAS File Source)

Reference Book

11 June 2014

Object Features: Point Cloud Features

365

Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

366

Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.8 Green
The Green image channel value associated with this point

34.8.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
Reference Book

11 June 2014

Object Features: Point Cloud Features

367

All not selected classes

Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.9 Blue
The Blueimage channel value associated with this point

34.9.1 Parameters
Image Layer (LAS File Source)

Reference Book

11 June 2014

Object Features: Point Cloud Features

368

Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

369

Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.10 Number of points

Formerly known as Counter. The Number of points that are mapped to the pixel.

34.10.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Returns

Select which return should be used for the conversion:

All
First
Last
You can dene to ignore elevation below the specied value.
Enter the value manually or choose:
Minimum allowed elevation

<none>
From Feature
From Array
You can dene to ignore elevation above the specied value.
Enter the value manually or choose:
Maximum allowed elevation

<none>
From Feature
From Array

Reference Book

11 June 2014

Object Features: Point Cloud Features

370

Point Filter - By Class

Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified value of red value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of green value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified value of blue value

Enter a value or select the mode for

ltering classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

Reference Book

11 June 2014

Object Features: Point Cloud Features

371

34.11 Class
Gives back information about the classication of the points. You can, for example,
choose Mode in the Result Mode to get information on the dominant class.

34.11.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

Ignore points with the specified green value

372
Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified blue value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

34.12 Distance to Camera

This feature calculates the distance to the camera. It can be used to apply different analysis approaches dependent on the distance to the camera or to exclude areas from the
analysis because of their distance. The distance unit is according to the unit of the point
cloud projection.

34.12.1 Parameters
Image Layer (LAS File Source)
Feature Parameters
Result Mode

Choose one of the following calculation algorithms:

Average
Standard Deviation
Minimum
Maximum
Reference Book

11 June 2014

Object Features: Point Cloud Features

373

Median
Mode
Sum
Quantile

Returns

Select which return should be used for the conversion:

All
First
Last
You can dene to ignore elevation below the specied value.
Enter the value manually or choose:
Minimum allowed elevation

<none>
From Feature
From Array
You can dene to ignore elevation above the specied value.
Enter the value manually or choose:
Maximum allowed elevation

<none>
From Feature
From Array
Point Filter - By Class
Filtering

Select the mode for ltering classes:

<none>
Selected classes
All not selected classes
Point Filter - By RGB value
Ignore points with the specified red value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Ignore points with the specified green value

Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes

Reference Book

11 June 2014

Object Features: Point Cloud Features

Ignore points with the specified blue value

374
Enter a value or select the mode for ltering

classes:
<none>
Selected classes
All not selected classes
Point Filter - By Distance

You can specify the minimum allowed input distance value

for calculation. Enter the value manually or choose:
Minimum allowed distance

<none>
From Feature
From Array
You can specify the maximum allowed input distance value
for calculation. Enter the value manually or choose:
Maximum allowed distance

<none>
From Feature
From Array

Reference Book

11 June 2014

35 Class-Related Features
Class-related features depend on image object features and refer to the class assigned to
image objects situated at any location in the image object hierarchy.
This location is specied for superobjects and sub-objects by the level distance dening
the vertical distance in the image object hierarchy. For neighbor image objects the location is specied by the spatial distance dening the horizontal distance in the image
object hierarchy. These feature distances can be edited.
Class-related features are global features because they are not related to individual image
objects. Class-Related Features are grouped as follows:
Relations to neighbor objects features are used to describe an image object by its
relationships to other image objects of a given class on the same image object level.
Relations to Sub-objects features describe an image object by its relationships to
other image objects of a given class on a lower image object level in the image
object hierarchy. You can use these features to evaluate sub-scale information because the resolution of image objects increases the lower you move in the image
object hierarchy.
Relations to Superobjects features describe an image object by its relations to other
image objects of a given class on a higher image object level in the image object
hierarchy. You can use these features to evaluate super-scale information because
the resolution of image objects decreases the higher you move in the image object
hierarchy.
Relations to Classication features are used to nd out about the current or potential classication of an image object.
Customized class-related features refer to class-related features. They are available
after they have been created.

35.1 Relations to Neighbor Objects

Class-Related Features > Relations to Neighbor Objects
Relations to neighbor objects features are used to describe an image object by its relationships to other image objects of a given class on the same image object level.

35.1.1 Existence of
Class-Related Features > Relations to Neighbor Objects > Existence Of

375

Class-Related Features

376

Existence of an image object assigned to a dened class in a certain perimeter (in pixels)
around the image object concerned. If an image object of the dened classication is
found within the perimeter, the feature value is 1 (= true), otherwise it would be 0 (= false).
The radius dening the perimeter can be determined by editing the feature distance.
Expression

0 if Nv (d, m) = , 1 if Nv (d, m) 6=
Feature Value Range

[0, 1]

35.1.2 Number Of
Class-Related Features > Relations to Neighbor Objects > Number Of
Number of objects belonging to the selected class in a certain distance (in pixels) around
the image object.
Parameters

v is the image object

d is the distance between neighbors
m is a class containing image objects
Expression

#Nv (d, m)
Feature Value Range

[0, ]

35.1.3 Border To
Class-Related Features > Relations to Neighbor Objects > Border To
The absolute border of an image object shared with neighboring objects of a dened
classication. If you use geo-referenced data, the feature value is the real border to
image objects of a dened class; otherwise it is the number of pixel edges shared with the
adjacent image objects, as by default the pixel edge-length is 1.
Parameters

b(v, u) is the length of common border between v and u

Nv (d) are neighbors to an image object v at a distance d

Reference Book

11 June 2014

Class-Related Features

377

Expression

b(v, u)

uNv (d,m)

Figure 35.1. The absolute border between unclassified and classified image objects

Feature Value Range

[0, ]

35.1.4 Rel. Border to

Class-Related Features > Relations to Neighbor Objects > Rel. Border To
The feature Rel. Border To determines the relative border length an object shares with
the image border. It describes the ratio of the shared border length of an image object
(with a neighboring image object assigned to a dened class) to the total border length. If
the relative border of an image object to image objects of a certain class is 1, the image
object is totally embedded in them.
If the relative border is 0.5 then the image object is surrounded by half of its border. The
relative border length can only be expressed in pixels.
Parameters

b(v, u) is the length of common border between v and u

Nv (d) are neighbors to an image object v at a distance d
bv is the image object border length
Expression

b(v, u)

uNv (d,m)

bv
Feature Value Range

[0, 1]

Reference Book

11 June 2014

Class-Related Features

378

Figure 35.2. Relative border between neighbors

Conditions

If the relative border is 0 then the class m does not exist. If the relative border is 1 then
the object v is completely surrounded by class m

35.1.5 Rel. Area of

Class-Related Features > Relations to Neighbor Objects > Rel. Area Of
The area covered by image objects of a selected class, found within a user-dened circular
area around the selected image object, divided by the total area of image objects inside
this area.
Editable Parameters

Class
Feature Distance: Radius (from the selected image object) of the area in pixels.
Parameters

Nv (d) are neighbors to an image object v at a distance d

#Pu is total number of pixels contained in Pu
Expression

#Pu

uNv (d,m)

#Pu

uNv (d)

Feature Value Range

[0, 1]

Reference Book

11 June 2014

Class-Related Features

379

Conditions

If the relative area is 0 then the class m does not exist. If the relative area is 1 then the
object v is completely surrounded by class m

35.1.6 Distance to
Class-Related Features > Relations to Neighbor Objects > Distance To
The distance (in pixels) of the image objects center concerned to the closest image objects center assigned to a dened class. The image objects on the line between the image
objects centers have to be of the dened class.
Parameters

d(v, u) is the distance between v and u

Vi (m) is the image object level of a class m
bv is the image object border length
Expression

min d(v, u)

uVi (m)

Figure 35.3. Distance between the centers of neighbors

Feature Value Range

[0, ]

35.1.7 Mean Diff. to

Class-Related Features > Relations to Neighbor Objects > Mean Diff. To
The mean difference of the layer L mean value of the image object concerned to the layer
L mean value of all image objects assigned to a dened class.

Reference Book

11 June 2014

Class-Related Features

380

Parameters

v is the image object

Nv (m) is the neighbor to an image object v of a class m
Expression

Nv (m))
(v,
Feature Value Range

[0, ]

35.1.8 Overlap of two Objects

Class-Related Features > Relations to Neighbor Objects > Overlap of two Objects
Calculates the overlap of image objects with image objects of a specied class. For more
information please refer also to Linking Operation Algorithms (p 120).
Parameters

Classes: Select the candidate classes - only objects of specied classes will be used
in the overlap calculation.
X-/Y-/Z-/T-position shift: a specied shift is applied to image objects pixels prior
to calculating the overlap e.g. relative shift in the x direction.
Overlap calculation
The following overlap calculation methods are available:
1. Do not use overlap: Overlap calculation is omitted. The parameter returns 1 if
there is any overlap with candidate image objects and 0 if there is no overlap.
2. Relative to larger object [0..1]: The ratio of the overlap area to the area of the larger
object (between candidate and current object) is calculated.
3. Relative to smaller object [0..1]: The ratio of the overlap area to the area of the
smaller object (between candidate and current object) is calculated.
4. Relative to current object [0..1]: The ratio of the overlap area to the area of the
current object is calculated.
5. Relative to candidate object [0..1]: The ratio of the overlap area to the area of the
candidate object is calculated.
6. Absolute [in pixels]: The number of overlapping pixels is calculated.
Feature Value Range

1.
2.
3.
4.
5.
6.

[0, 1]
[0..1]
[0..1]
[0..1]
[0..1]
[0, ]

Reference Book

11 June 2014

Class-Related Features

381

35.2 Relations to Sub-Objects

Class-Related Features > Relations to Sub Objects
Relations to sub-objects features describe an image object by its relationships to other
image objects of a given class on a lower image object level in the image object hierarchy.
You can use these features to evaluate sub-scale information because the resolution of
image objects increases the lower you move in the image object hierarchy.

35.2.1 Existence Of
Class-Related Features > Relations to Sub Objects > Existence Of
The Existence Of feature checks if there is at least one sub-object assigned to a dened
class. If there is one, the feature value is 1 (= true), otherwise the feature value is 0 (=
false).
Parameters

v is the image object

d is the distance between neighbors
m is a class containing image objects
Expression

0 if Sv (d, m) = , 1 if Sv (d, m) 6=
Feature Value Range

[0, 1]

35.2.2 Number of
Class-Related Features > Relations to Sub Objects > Number Of
The number of sub-objects assigned to a dened class.
Parameters

v is the image object

d is the distance between neighbors
m is a class containing image objects
Expression

#Sv (d, m)

Reference Book

11 June 2014

Class-Related Features

382

Feature Value Range

[0, ]

35.2.3 Area of
Class-Related Features > Relations to Sub Objects > Area Of
The absolute area covered by sub-objects assigned to a dened class. If your data are
geo-referenced, the feature value represents the real area.
Parameters

d is the distance
m is the class
M is a sub-object in class m
Expression

#PM

MESv (d,m)

Feature Value Range

[0, ]

35.2.4 Rel. Area of

Class-Related Features > Relations to Sub Objects > Rel. Area Of
Area covered by sub-objects assigned to a given class divided by the total area of the
image object concerned.
Editable Parameters

Class
Image object level Distance: Downward distance of image object levels in the
image object hierarchy between the image object and the sub-object.
Parameters

d is the image object level distance

m is the class
M is a sub-object in class m

Reference Book

11 June 2014

Class-Related Features

383

Expression

#PM

MESv (d,m)

#Pv
Feature Value Range

[0, 1]

35.2.5 Clark Aggregation Index

Class-Related Features > Relations to Sub Objects > Clark Aggregation Index
For a superobject the Clark aggregation index gives evidence about the spatial distribution
of its sub-objects of a certain class.
Parameters

D(x) is the mean spatial distance to next neighbor of the sub-objects of the class x
N(x) is the number of sub-objects of class x
A is the number of pixels of the superobject (Area)
Obs_mean_dist is the observed mean distance of sub objects to their spatial nearest
neighbor with Obs_mean_dist = (D(x))
N(x)
Exp_mean_dist is the expected mean distance of sub objects to their spatial nearest
1
neighbor with Exp_mean_dist = q
2 N(x)
A

CAI is the Clark aggregation index

Expression

CAI =

Obs_Mean_Distance
Exp_Mean_Distance

Feature Value Range

[0, 2.149]
0 heavily clumped sub-objects
1 homogeneous spatial distribution of sub-objects
2.149 hexagonal distribution (edges of a honeycomb) of the sub-objects

35.3 Relations to Superobjects

Class-Related Features > Relations to Super Objects
Relations to superobjects features describe an image object by its relations to other image
objects of a given class on a higher image object level in the image object hierarchy. You
Reference Book

11 June 2014

Class-Related Features

384

can use these features to evaluate super-scale information because the resolution of image
objects decreases the higher you move in the image object hierarchy.

35.3.1 Existence of
Class-Related Features > Relations to Super Objects > Existence of
Checks if the superobject is assigned to a dened class. If this is true, the feature value is
1, otherwise 0.
Parameters

v is the image object

d is the distance to superobjects
m is the a class containing image objects
Expression

0 if Uv (d, m) = , 1 if Uv (d, m) 6=
Feature Value Range

[0, 1]

35.4 Relations to Classification

Class-Related Features > Relations to Classification
Relations to classication features are used to nd out about the current or potential
classication of an image object.

35.4.1 Membership to
Class-Related Features > Relations to Classification > Membership to
In some cases it is important to incorporate the membership value to different classes in
one class. This function allows explicit addressing of the membership values to different
classes. If the membership value is below the assignment threshold, this value turns to 0.
Parameters

v is the image object

m is the a class containing image objects
m) stored membership value of an image object v to a class m
(v,

Reference Book

11 June 2014

Class-Related Features

385

Expression

m)
(v,
Feature Value Range

[0, 1]

35.4.2 Classified as
Class-Related Features > Relations to Classification > Classified As
The idea of this feature is to enable the user to refer to the classication of an image
object without regard to the membership value. It can be used to freeze a classication.
Parameters

v is the image object

m is a class containing image objects
Expression

m(v)
Feature Value Range

[0, 1]

35.4.3 Classification Value of

Class-Related Features > Relations to Classification > Classification Value Of
This feature Classication value of allows you to explicitly address the membership values to all classes. As opposed to the feature Membership to it is possible to apply all
membership values to all classes without restrictions.
Parameters

v is the image object

m is the a class containing image objects
m) is a fuzzy membership value of an image object v to a class m
(v,
Expression

(v, m)

Reference Book

11 June 2014

Class-Related Features

386

Feature Value Range

[0, 1]

35.4.4 Class Name

Class-Related Features > Relations to Classification > Class Name
The Class name feature returns the name of the class (or superclass) of an image object
(or its superobject).
Parameters

Distance in class hierarchy species the number of hierarchical levels when navigating
from class to superclass. Using a distance of 0 the class name is returned, a distance of
1 will return the superclass name and so on. Distance in image object hierarchy species
the number of hierarchical levels when navigating from object to superobject. Using a
distance of 0 the class of the image object is used as a starting point for the navigation in
the class hierarchy, a distance of 1 will start at the class of the superobject.

35.4.5 Class Color

Class-Related Features > Relations to Classification > Class Color
The Class color feature returns either the red, green or blue color component of the class
(or superclass) of an image object (or its superobject).
Parameters

Color component is red, green or blue.

Distance in class hierarchy species the number of hierarchical levels when navigating
from class to superclass. Using a distance of 0 the class name is returned, a distance of 1
will return the superclass name and so on.
Distance in image object hierarchy species the number of hierarchical levels when navigating from object to superobject. Using a distance of 0 the class of the image object is
used as a starting point for the navigation in the class hierarchy, a distance of 1 will start
at the class of the superobject.

35.4.6 Assigned Class

Class-Related Features > Relations to Classification > Assigned Class
Based on this feature you can assign class names from the image object domain to a class
array.

Reference Book

11 June 2014

Class-Related Features

387

Process Example

In the algorithm update array select the Algorithm parameter > Feature > Class-Related
features > Relations to classication > Assigned class This results for instance in the
following process: at Main Level: update array class array: add from feature Assigned
class
Project Example

Please refer to the eCognition community for two projects Array Example 1 & 2 using
this feature: http://community.ecognition.com/home/Arrays%20Example%20%231.zip
/view http://community.ecognition.com/home/Arrays%20Example%20%232.zip/view
(To search the corresponding process press the shortcut Ctrl + F in eCognition or go to
the menu Process > Find and Replace and insert in the eld Name: assigned class)

Reference Book

11 June 2014

36 Linked Object Features

Linked Object features are calculated by evaluating linked objects themselves.

36.1 Linked Object Count

Linked Object Features > Linked Objects Count
The number of objects connected by object links fullling the lter parameters.
This feature traverses all links complying with the lter parameters and counts all objects
visited while traversing. You can use this feature to measure the number of objects in a
linked structure.

36.1.1 Editable Parameters

Link Class Filter: Select the classes that contain the links to be counted
Link Direction: Filter on the link direction:
All: Count all links, independent of link direction.
In: Count only inbound links.
Out: Count only outbound links.
Max. Distance: Maximum number of links in a path that will be taken into account.
This parameter lets you limit the counting to an objects local neighborhood. Use
0 for objects one link away, 1 for objects two links away and 999999 to count all
linked objects.

36.2 Linked Objects Statistics

Linked Object Features > Linked Objects Statistics
This feature allows computing a variety of statistical values for a group of linked objects.

36.2.1 Editable Parameters

Operation: Select the type of statistical operation to be performed:
Sum: Compute the sum of the selected feature for all linked objects
Mean: Compute the mean value of the selected feature for all linked objects

388

Linked Object Features

389

Figure 36.1. Example of a linked objects domain of outgoing direction within a maximum

distance of 3

Std Dev: Compute the standard deviation of the selected feature for all linked
objects
Min: Returns the minimum value of the selected feature for all linked objects
Max: Returns the maximum value of the selected feature for all linked objects
Is Min: Returns 1 if the image object is the minimum for the selected feature
for all linked objects
Is Max: Returns 1 if the image object is the maximum for the selected feature
for all linked objects
Count: Counts the linked objects. Ignores the selected feature
Feature: Select the image object feature that will be used for the statistical operation.
Link Class Filter: Select the classes that contain the links to be counted
Candidate Condition: Select the condition the linked objects must fulll to be
included in the statistical computation
Link Direction: Filter on the link direction:
All: Use all links, independent of link direction.
In: Use only inbound links.
Out: Use only outbound links.
Max. Distance: The maximum number of links in a path that will be taken into
account. This parameter allows you to limit the counting to the local neighborhood
of the object. Use 0 for objects one link away, 1 for objects two links away and
999999 to count all linked objects

36.3 Linked Weight to PPO

Linked Object Features > Linked Weight to PPO
This feature calculates the link weight to a selected PPO. The weight is a measure of the
degree of the overlap: if the overlap of the seed and target objects is 100% then the weight
is 1. If the overlap is 50% only, then the weight is 0.5. If there are many links available,
this should return the maximum of the given link weights.

Reference Book

11 June 2014

Linked Object Features

390

In order to display meaningful values, create a feature with default values, select the
image object of interested using Ctrl + left-mouse click (the object will outlined in dark
green), then select a linked image object.

36.3.1 Editable Parameters

Hierarchical Distance to PPO (the default is 1)
Link Class Filter: Select the classes that contain the links to be taken into account.
Link Direction: Filter on the link direction:
All: Use all links, independent of link direction
In: Use only inbound links
Out: Use only outbound links.
Hierarchical Distance to PPO

The number of hierarchical levels to reach PPO.

Link Class Filter

Restricts the active link classes.

Link Direction

Link Direction species the permitted link directions that create the linked object domain.

Reference Book

11 June 2014

37 Scene Features
Scene features return properties referring to the entire scene or map. They are global
because they are not related to individual image objects. Scene Features are grouped as
follows:
Scene variables are global variables that exist only once within a project. They are
independent of the current image object.
Class-related scene features provide information on all image objects of a given
class per map.
Scene-related features provide information on the scene.
Customized scene features refer to scene features. They are available after they
have been created.

37.1 Scene Variables

Scene Features > Scene Variables
Scene Variables are global variables that exist only once within a project. They are independent of the current image object. If they exist, all scene variables are listed in
the feature tree. You can dene a new scene variable by clicking on Create New Scene
Variable to store global interim values.

37.1.1 Editable Parameters

Name
Value: Insert an initial value for the variable
Choose whether the new variable is numeric (Double) or textual (String)
Shared: Select if you want to share the new variable among different rule sets.

37.2 Class-Related
Scene Features > Class-Related
Class-related scene features provide information on all image objects of a given class per
map.

391

Scene Features

392

37.2.1 Number of Classified Objects

Scene Features > Class-Related > Number of Classified Objects
The absolute number of all image objects of the selected class on all image object levels.
Editable Parameters

Class
Parameters

V (m) is all image objects classied as m

m is the given class
Expression

#V (m)
Feature Value Range

[0, number of image objects]

37.2.2 Number of Samples Per Class

Scene Features > Class-Related > Number of Samples Per Class
The number of all samples of the selected class on all image object levels.
Editable Parameters

Class
Feature Value Range

[0, number of samples]

37.2.3 Area of Classified Objects

Scene Features > Class-Related > Area of Classified Objects
Editable Parameters

Class

Reference Book

11 June 2014

Scene Features

393

Parameters

v is the image object

m is the given class
V (m) is all image objects classied as m
#Pv is the total number of pixels of image object v

Expression

#Pv

veV (m)

Feature Value Range

[0, sx sy sz st]

37.2.4 Layer Mean of Classified Objects

Scene Features > Class-Related > Layer Mean of Classified Objects
The mean of all image objects of the selected class on the selected image object levels.
Editable Parameters

Class
Image layer
Parameters

v is the image object

m is the given class
V (m) is all image objects classied as m
ck
v is the mean intensity layer of an image object v

Expression

1
ck (v)
#V (m) veV(m)
Feature Value Range

[0, 1]

37.2.5 Layer Std. Dev. of Classified Objects

Scene Features > Class-Related > Layer Stddev. of Classified Objects

Reference Book

11 June 2014

Scene Features

394

Editable Parameters

Class
Image layer
Parameters

v is the image object

m is the given class
V (m) is all image objects classied as m
ck
v is the mean intensity layer of an image object v

Expression

v
u
u
k V (m) = t

1
#V (m)

veV (m)

2
ck (v)

1
#V (m)

veV (m)

2 !
ck (v)

Feature Value Range

[0, 1]

37.2.6 Statistic of Object Value

Scene Features > Class-Related > Statistic of Object Value
This feature allows the calculation of statistical operators over an existing objects feature.
Editable Parameters

Class
Statistical Operation
Feature

37.2.7 Class Variables

Scene Features > Class-Related > Class Variable
Class Variables use classes as values. In a rule set they can be used instead of ordinary
classes to which they point. If any exist, all variables that use classes as values are listed
in the feature tree. You can dene a new class variable by clicking on Create new Class
Variable.
Editable Parameters

Name: To dene an new class variable, enter a name and click OK. The Create
Class Variable dialog box opens for class variable settings.

Reference Book

11 June 2014

Scene Features

395

37.3 Scene-Related
Scene Features > Scene-Related
Scene-related features provide information on the scene.

37.3.1 Existence of Object Level

Scene Features > Scene-Related > Existence of Object Level
This feature lets you test the existence of an object level with a given name. If the object
level with the given name does not exist within the project, the feature value is 1 (true),
otherwise it is 0 (false).
Editable Parameters

Level Name
Feature Value Range

[0, 1]

37.3.2 Existence of Image Layer

Scene Features > Scene-Related > Existence of Image Layer
Existence of a given image layer.
If the image layer with the given alias exists within the map the feature value is 1 (= true),
otherwise it is 0 (= false).
Editable Parameter

Image Layer alias

Feature Value Range

[0, 1]

37.3.3 Existence of Thematic Layer

Scene Features > Scene-Related > Existence of Thematic Layer
Existence of a given thematic layer. If the thematic layer with the given alias exists within
the map the feature value is 1 (true), otherwise it is 0 (false).

Reference Book

11 June 2014

Scene Features

396

Editable Parameter

Thematic layer alias

Feature Value Range

[0, 1]

37.3.4 Existence of Map

Scene Features > Scene-Related > Existence of Map
This feature lets you test the existence of a map with a given name.
If the map with the given name does exist within the project, the feature value is 1 (true),
otherwise it is 0 (false).
Editable Parameter

Map
Feature Value Range

[0, 1]

37.3.5 Mean of Scene

Scene Features > Scene-Related > Mean of Scene
Mean value of the selected image layer.
Editable Parameters

Image Layer
Expression

ck
Feature Value Range
max
[cmin
k , ck ]

Reference Book

11 June 2014

Scene Features

397

37.3.6 Standard deviation

Scene Features > Scene-Related > Std Dev
Standard deviation of the selected layer.
Expression

37.3.7 Smallest Actual Pixel Value

Scene Features > Scene-Related > Smallest Actual Pixel Value
Darkest actual intensity value of all pixel values of a given image layer.
Editable Parameter

Image Layer
Expression
min

c k

Feature Value Range

max
[cmin
k , ck ]

37.3.8 Largest Actual Pixel Value

Scene Features > Scene-Related > Largest Actual Pixel Value
Brightest actual intensity value of all pixel values of a given image layer.
Editable Parameter

Image Layer
Expression
max

c k

Feature Value Range

max
[cmin
k , ck ]

Reference Book

11 June 2014

Scene Features

398

37.3.9 Validity of Region

Scene Features > Scene-Related > Validity of Region
The Validity of region feature checks if a given region is located within the scene.
If the given regions is located within the scene the feature value is 1 (true), otherwise it is
0 (false).
Editable Parameter

Region
Feature Value Range

[0, 1]

37.3.10 Mode
Scene Features > Scene-Related > Mode
Gives back the mean value of the objects for a dened layer, which repeats most often
per scene. Example: A scene has the following object mean values: 1 2222 333 444444
55 4 would be the mode, because it appears most often
There is the possibility, that values apprears equally often. Like: 11 2222 3333 4444 55,
then you can choose from the parameters, whether you want to get back the mimimum
( smallest, here: 4) the maximum ( largest, here: 5) or the median (here: 3)
Editable Parameters

Image Layer; Return result Choose one of the following calculations for Return result
Minimum
Maximum
Median

37.3.11 Quantile
Scene Features > Scene-Related > Quantile
Gives back the dened quantile of a dened layer for the scene.
Calculation: rst values are sorted from small to large. Associates the ordered values with
equally spaced sample fractions. It reads out the value at the dened quantile position.
Editable Parameters

Select a layer from the Layer drop-down box, from which to extract the quantile
Insert a percentage value into the Quantile eld (or get the value from a feature or
array item).

Reference Book

11 June 2014

Scene Features

399

Parameters

The quantile is given by the element of the list Y with index equal to int(r)

Y is the list of pixel intensities (sorted in increasing order)

N is the number of pixels
p is a number between 0 and 100 (i.e. a percentage)
p
N
The rank, r = 100

Feature Value Range

[min layer value, max layer value]

37.3.12 Active Pixel t-value

Scene Features > Scene-Related > Active pixel t-value
The pixel/voxel t -coordinate of the last-clicked cursor position.
Feature Value Range

[0, scene extent]

37.3.13 Active Pixel x-value

Scene Features > Scene-Related > Active pixel x-value
The pixel x -coordinate of the last-clicked cursor position.
Feature Value Range

[0, scene extent]

37.3.14 Active Pixel y-value

Scene Features > Scene-Related > Active pixel y-value
The pixel y -coordinate of the last-clicked cursor position.
Feature Value Range

[0, scene extent]

Reference Book

11 June 2014

Scene Features

400

37.3.15 Active Pixel z-value

Scene Features > Scene-Related > Active pixel z-value
The pixel/voxel z -coordinate of the last-clicked cursor position.
Feature Value Range

[0, scene extent]

37.3.16 Image Directory

Scene Features > Scene-Related > Image Directory
Returns the directory folder name where image les are located. In case more than one
image is loaded the directory of the rst image is returned. Example: ImageAnalysis

37.3.17 Image Full Path

Scene Features > Scene-Related > Image Full Path
Returns the full path of loaded image les including the image le name. In case
more than one image is loaded the path of the rst image is returned. Example:
\\ImageServer\Data\Forest\ImageAnalysis\Trees1.tif

37.3.18 Image Name

Scene Features > Scene-Related > Image Name
Returns the name of loaded image le including the image le name. In case more than
one image is loaded the name of the rst image is returned. Example: Trees1.tif

37.3.19 Is Active Vector Layer Changed

Scene Features > Scene-Related > Is Active Vector Layer Changed
Returns a value if a user has edited a polygon (a value of one is true).
Feature Value Range

[0, 1]

37.3.20 Map Origin T

Scene Features > Scene-Related > Map Origin T
The t -co-ordinate of the origin of a map. If the unit is a pixel, the origin is 0; if it is a
co-ordinate, the origin is the absolute position of the map.

Reference Book

11 June 2014

Scene Features

401

37.3.21 Map Origin X

Scene Features > Scene-Related > Map Origin X
The x -co-ordinate of the origin of a map. If the unit is a pixel, the origin is 0; if it is a
co-ordinate, the origin is the absolute position of the map.

37.3.22 Map Origin Y

Scene Features > Scene-Related > Map Origin Y
The y -co-ordinate of the origin of a map. If the unit is a pixel, the origin is 0; if it is a
co-ordinate, the origin is the absolute position of the map.

37.3.23 Map Origin Z

Scene Features > Scene-Related > Map Origin Z
The z -co-ordinate of the origin of a map. If the unit is a pixel, the origin is 0; if it is a
co-ordinate, the origin is the absolute position of the map.

37.3.24 Map Size T

Scene Features > Scene-Related > Map Size T
The extent of a scene in time, calculated by the number of frames multiplied by the
temporal distance between slices in time co-ordinate units.
Parameters

st is the extent of a scene s in the t -direction

uframes is the temporal distance between slices in the time co-ordinate unit
Expression

st = #(frames) uframes
Feature Value Range

[1, ]

37.3.25 Map Size X

Scene Features > Scene-Related > Map Size X
The horizontal extent of a scene in the display unit.

Reference Book

11 June 2014

Scene Features

402

Parameters

sx is the extent of a scene s in the x -direction

# (pixels)x u is number of pixels in the x -direction
u is the size of a pixel in the co-ordinate system unit
Expression

sx = #(pixels)x u
Feature Value Range

[1, ]

37.3.26 Map Size Y

Scene Features > Scene-Related > Map Size Y
The horizontal extent of a scene in the display unit.
Parameters

sy is the extent of a scene s in the y -direction

# (pixels)y u is number of pixels in the y -direction
u is the size of a pixel in the co-ordinate system unit
Expression

sy = #(pixels)y u
Feature Value Range

[1, ]

37.3.27 Map Size Z

Scene Features > Scene-Related > Map Size Z
The extent of a scene in depth, calculated by the number of slices multiplied by the spatial
distance between slices in the co-ordinate system units.
Parameters

sx is the extent of a scene s in the x -direction

{$u_{\mathrm {slices}} is the spatial distance between slices in the co-ordinate
system unit

Reference Book

11 June 2014

Scene Features

403

Expression

sx = #(slices)y uslices
Feature Value Range

[1, ]

37.3.28 Number of Image Layers

Scene Features > Scene-Related > Number of Image Layers
Number of layers K of a scene.

37.3.29 Number of Maps

Scene Features > Scene-Related > Number of Maps
Number of existing maps.

37.3.30 Number of Objects

Scene Features > Scene-Related > Number of Objects
Number of image objects of any class on all image object levels of the scene including
unclassied image objects.
Expression

#V
Feature Value Range

[0, 1]

37.3.31 Number of Pixels in Scene

Scene Features > Scene-Related > Number of Pixels in Scene
Number of pixels in the pixel layer of the scene.

Reference Book

11 June 2014

Scene Features

404

Parameters

sx is the scene extent in x

sy is the scene extent in y
sz is the scene extent in z
st is the scene extent in t
(sx, sy, sz, st) is the scene extent

Expression

sx sy sz st
Feature Value Range

[0, scene extent]

37.3.32 Number of Samples

Scene Features > Scene-Related > Number of Samples
Number of all samples on all image object levels of the map.
Feature Value Range

[0, number of samples]

37.3.33 Number of Thematic Layers

Scene Features > Scene-Related > Number of Thematic Layers
The number of thematic layers layers included in the map.

37.3.34 Original Scene ID

Scene Features > Scene-Related > Original Scene ID
The original identication number of the scene. It is generated automatically by eCognition Developer 9.0.

37.3.35 Original Scene Name

Scene Features > Scene-Related > Original Scene Name
The original name of the scene. It is generated automatically by eCognition Developer
9.0.

Reference Book

11 June 2014

Scene Features

405

37.3.36 Scene ID
Scene Features > Scene-Related > Scene ID
The identication number of the scene. It is generated automatically by eCognition Developer 9.0.

37.3.37 Scene Magnification

Scene Features > Scene-Related > Scene Magnification
Returns the scene magnication dened in a rule set.

37.3.38 Scene Name

Scene Features > Scene-Related > Scene Name
Returns the scene name. The scene name differs from the image name if you are working
with workspaces. Example: tile000

37.3.39 Scene Pixel Size

Scene Features > Scene-Related > Scene Pixel Size
The scene unit that corresponds to one pixel, as dened in the metadata or in the Modify
Project dialog box.

37.3.40 Scene Resolution

Scene Features > Scene-Related > Scene Resolution
The resolution of the scene as given in the metadata of the project. The resulting number
represents the size of a pixel in co-ordinate system unit. The value is 1 if no resolution is
set.

37.3.41 Scene Bit Depth

Scene Features > Scene-Related > Scene Bit Depth
Returns a value if the program has read a bits per sample value in the metadata.

37.3.42 Second Level Scene Name

Scene Features > Scene-Related > Second Level Scene Name
Displays the name of second-level scenes.

Reference Book

11 June 2014

Scene Features

406

37.3.43 Slice Distance

Scene Features > Scene-Related > Slice Distance
The spatial distance between slices in the co-ordinate system unit.
Expression

uslices

37.3.44 Time Series Distance

Scene Features > Scene-Related > Time Series Distance
The temporal distance between slices in the time co-ordinate unit.
Expression

uframes

37.3.45 Top Scene ID

Scene Features > Scene-Related > Top Scene ID
The identication number of the top scene.
A top scene is the original scene that was used to create scene copies, subsets, or tiles.

37.3.46 Top Scene Name

Scene Features > Scene-Related > Top Scene ID
The name of the top scene.
A top scene is the original scene that was used to create scene copies, subsets, or tiles.

37.3.47 User Name

Scene Features > Scene-Related > User Name
The user name.

37.3.48 Workspace Full Path

Scene Features > Scene-Related > Workspace Full Path
Returns the full path of the workspace including the workspace name.
C:\Users\EClark\AppData\Roaming\eCognition\Data\Trees1.dpj

Reference Book

Example:

11 June 2014

Scene Features

407

37.4 Rule-Set Related

37.4.1 Rule set items
Scene Features > Rule set-Related > Rule set items
These features gives you back lists to dynamically ll arrays, i.e. to initialize arrays.

All existing classes

All existing image layers
All existing image object levels
All existing maps
All existing temporary image layers
All existing thematic layers

37.4.2 Class Variable

Scene Features > Rule-Set Related > Class Variable
Enter a name, color and value.

37.4.3 Level Variable

Scene Features > Rule-Set Related > Level Variable
Enter a name and value.

37.4.4 Map Name Variable

Scene Features > Rule-Set Related > Map Name Variable
Enter a name and value.

37.4.5 Image Layer Variable

Scene Features > Rule-Set Related > Image Layer Variable
Enter a name and value.

37.4.6 Thematic Layer Variable

Scene Features > Rule-Set Related > Thematic Layer Variable
Enter a name and value.

Reference Book

11 June 2014

Scene Features

408

37.4.7 Number of Features in Feature List

Scene Features > Rule-Set Related > Number of Features in Feature List
Enter a name and feature(s).

37.4.8 Number of Specified Features in Feature List

Scene Features > Rule-Set Related > Number of Specified Features in Feature List
Enter a name and feature(s) for the Feature List and Feature(s).

37.4.9 Rule Set Array Values

Scene Features > Rule-Set Related > Rule Set Array Values
Enter the name of an array.

37.4.10 Rule Set Array Size

Scene Features > Rule-Set Related > Rule Set Array Size
Enter the name of an array.

37.4.11 Rule Set Array Item

Scene Features > Rule-Set Related > Rule Set Array Item
Enter the name of an array and the index value to access an array item.

37.5 Architect Related

Scene Features > Rule-Set Related

37.5.1 Number of Actions

Scene Features > Rule-Set Related > Number of Actions
The Number of Actions features allows a rule set to keep track of the structure of the
analysis.
If the value is 0, the action is not present in the library
If the value is 1, the action is present in the library
Where there are multiple actions, the value increases.

Reference Book

11 June 2014

Scene Features

409

37.6 File System

Scene Features > File System

37.6.1 Resolve Path

Scene Features > File System > Resolve Path
Resolve Path lets you enter a string that will be converted to a le path relative to the
Workspace folder. Paths can contain shortcuts and variables, which will be converted to
the le path.
Supported Shortcuts

eCognition Developer 9.0 will replace the following variables with actual values:
Shortcut

Description

"{:Workspc.Name}"

Name of the workspace

"{:Workspc.Guid}"

GUID of the workspace

"{:Workspc.Dir}"

Path of the workspace le

"{:Project.Name}"

Name of the scene (absolute in the workspace)

"{:Project.Guid}"

GUID of the scene

"{:Scene.Dir}"

Directory of the scene

"{:Scene.Name}"

Name of the scene

"{:ImgLayer(n).Dir}"

Path to the input image layer n

"{:Workspc.OutputRoot}" Path of the output root

"{:Workspc.InputRoot}"

Path of the input root

"{:ActionLib.Dir}"

Directory of the action library (if any)

"{:Variable:abcde}"

Value of variable abcde

"{:Application.Dir}"

Directory of the loaded application/portal (if any)

Example

The Workspace le is located at C:\My Samples\Workspaces\1206\WKSP_201101111206.dpj

and the value of the variable str_temp1 is 91.32.12. The Resolve Path feature will convert
the input path
{:Workspc.OutputRoot}\results\report_Variable [str_temp1].html to
C:\My Samples\Workspaces\WKSP_201101111206\results\report_91.32.12.html

37.7 UI Related
Scene Features > UI Related

Reference Book

11 June 2014

Scene Features

410

37.7.1 Equalization
Scene Features > UI Related > Equalization
Equalization features return the level equalization values selected by users (which can be
changed via the Edit Image Layer Mixing dialog box, or by dragging the mouse across
the image with the right mouse button, if this option is enabled in Tools > Options).
Equalization Mode

Values may be:

Linear
Linear Inverse
Gamma Correction (Positive)
Gamma Correction (Negative)
Gamma Correction (Positive) Inverse
Gamma Correction (Negative) Inverse

Window Level Center

The mid-point of the dened equalization range.

Window Level Max

The end-point of the dened equalization range.

Window Level Min

The beginning of the dened equalization range.

Window Level Width

The equalization range either side of the center (Window Level Max minus Window Level
Min)

Reference Book

11 June 2014

38 Process Related Features

Process-related features are image object dependent features. They involve the relationship of a child process image object to a parent process. They are used in local processing.
A process-related feature refers to a relation of an image objects to a parent process object
(PPO) of a given process distance in the process hierarchy.
The process distance in the process hierarchy represents the upward distance of hierarchical levels in process tree between a process and the parent process. The process distance
is a basic parameter of process-related features. Practically, that is the number of hierarchy levels in the Process Tree window above the current editing line, in which you nd
the denition of the parent object.
In the Process Tree, hierarchical levels are indicated by different indentations.

Figure 38.1. Process Tree window displaying a prototype of a process hierarchy. The pro-

cesses are named according to their connection mode

For example, a process distance of 1 means that the parent process is located one hierarchical level above the current process. A process distance of 2 means that the parent
process is located two hierarchical levels above the current process. Figuratively, you
may say, a process distance of 2 denes the grandparent process.

38.1 Customized
Process-Related Features > Customized

38.1.1 Diff. PPO

Process-Related Features > Customized > Create New Diff. PPO
411

Process Related Features

412

Dene a customized feature in relation to the difference between a given feature value of
an image object and the feature value of its parent process object (PPO).
Editable Parameters

Distance: Upward distance of hierarchical levels in process tree between a process

and the parent process.
Feature
Parameters

v is the image object

f is any feature
is a parent process object (PPO)
Expression

f (v) f ()
Feature Value Range

The range depends on the value of the feature in use.

Conditions

If f () = 0 the formula is undened.

38.1.2 Ratio PPO

Process-Related Features > Customized > Ratio PPO
Dene a customized feature by the ratio between a given feature value of an image object
and the feature value of its parent process object (PPO).
Editable Parameters

Distance: Upward distance of hierarchical levels in process tree between a process

and the parent process.
Feature
Parameters

v is the image object

f is any feature
is a parent process object (PPO)

Reference Book

11 June 2014

Process Related Features

413

Expression

f (v)
f ()
Feature Value Range

The range depends on the value of the feature in use.

Conditions

If f () = 0 the formula is undened.

38.2 Border to PPO

Process-Related Features > Border to PPO
The absolute border of an image object shared with its parent process object.

38.2.1 Editable Parameters

Distance: Upward distance of hierarchical levels in process tree between a process
and the parent process.

38.2.2 Parameters
b(v, ) is the length of common border between v and with the parent process
object

38.2.3 Expression
b(v, r)

38.2.4 Feature Value Range

[0, maximum size]

38.3 Distance to PPO

The distance between two parent process objects.

Reference Book

11 June 2014

Process Related Features

414

38.3.1 Editable Parameter

Distance: Upward distance of hierarchical levels in process tree between a process
and the parent process.

38.4 Elliptic Dist. from PPO

Process-Related Features > Elliptic Dist. from PPO
The elliptic distance of an image object to its parent process object (PPO).
The ellipse is derived from area and center of gravity of the parent process object. The
feature value is the distance between the center of gravity of the image object and the
ellipse.

38.4.1 Parameters
xv , yv

38.4.2 Editable Parameters

Metadata Name: Name of the metadata source as used in the Modify Project dialog box. This name will be used in the feature tree for grouping corresponding
metadata items. You can type a name of a not existing metadata source, to create a
feature group in advance.
Name: Name of the metadata item as used in the source data.
Type: Select the type of the metadata item: string, double, or integer.
Distance: Upward distance of hierarchical levels in process tree between a process
and the parent process.

38.4.3 Expression
(xv , yv )

38.4.4 Feature Value Range

[0, ]

38.5 Rel. border to PPO

Process-Related Features > Rel. Border to PPO
The ratio of the border length of an image object shared with the parent process object
(PPO) to its total border length.

Reference Book

11 June 2014

Process Related Features

415

38.5.1 Editable Parameters

Process Distance: Upward distance of hierarchical levels in process tree between a process and the parent process.

38.5.2 Parameters
bv is the image object border length
b(v, ) is the length of common border between v and with the parent process
object

38.5.3 Expression
b(v, )
bv

38.5.4 Feature Value Range

[0, 1]

38.6 Same Superobject as PPO

Process-Related Features > Same Super Object as PPO
The Same super object as PPO feature checks whether an image object and its parent
process object (PPO) are parts of the same superobject.

38.6.1 Editable Parameters

Process Distance: Upward distance of hierarchical levels in process tree between a
process and the parent process.

38.6.2 Parameters
v is the image object
is the parent process object (PPO)
Uv (d) is the superobject of an image object v at a distance d

38.6.3 Expression
1 : Uv (d) = U (d)
0 : Uv (d) 6= U (d)

Reference Book

11 June 2014

Process Related Features

416

38.6.4 Feature Value Range

[0, 1]

38.7 Series ID
Process-Related Features > Series ID
The Execute Child as Series algorithm executes its child domains based on the Number of
Cycles parameter. For each execution, a unique identier is generated and the parameter
Series Name prexes this identier. Therefore, if the domain has four image objects, and
the number of cycles is set to 1, the child processes will be executed four times (once for
each image object). On each cycle, the value attached to the series increases by one, for
example:
1st cycle = series1
2nd cycle = series2
Therefore, the feature acts as a loop counter inside the Execute Child as Series algorithm.

Reference Book

11 June 2014

39 Region Features
Region features return properties referring to a given region. They are global because
they are not related to individual image objects. They are grouped as follows:
Region-related features provide information on a given region.
Layer-related region features evaluate the rst and second statistical moment
(mean, standard deviation) of a regions pixel value.
Class-related region features provide information on all image objects of a given
class per region.

39.1 Region-Related
Region Features > Region-Related
Region-related features provide information on a given region.

39.1.1 Number of Pixels in Region

Region Features > Region-Related > Number of pixels in region
Number of pixel, voxels, or voxel series forming a region.
Editable Parameters

Region
Parameters

Rx is the region extent in x

Ry is the region extent in y
Rz is the region extent in z
Rt is the region extent in t

Expression

Rx Ry Rz Rt

417

Region Features

418

Feature Value Range

[0, ]

39.1.2 T Extent
Region Features > Region Related > T Extent
Number of frames within a region.
Editable Parameters

Region
Expression

Rt
Feature Value Range

[0, ]

39.1.3 T Origin
Region Features > Region Related > T Origin
The t -co-ordinate of the origin of a region.
Editable Parameters

Region
Expression

tG
Feature Value Range

[0, ]

39.1.4 X Extent
Region Features > Region Related > X Extent
The extent of a region in x

Reference Book

11 June 2014

Region Features

419

Editable Parameters

Region
Expression

Rx
Feature Value Range

[0, ]

39.1.5 X Origin
Region Features > Region Related > X Origin
The x -co-ordinate of the origin of a region.

39.1.6 Y Extent
Region Features > Region Related > Y Extent
The pixel extent of a region in y .
Editable Parameters

Region
Expression

xG
Feature Value Range

[0, ]

39.1.7 Y Origin
Region Features > Region Related > Y Origin
The y -co-ordinate of the origin of a region.
Editable Parameters

Region

Reference Book

11 June 2014

Region Features

420

Expression

yG
Feature Value Range

[0, ]

39.1.8 Z Extent
Region Features > Region Related > Z Extent
The number of slices within a region.
Editable Parameters

Region
Expression

Rz
Feature Value Range

[0, ]

39.1.9 Z Origin
Region Features > Region Related > Z Origin
The z -co-ordinate of the origin of a region.
Editable Parameter

Region
Expression

zG
Feature Value Range

[0, ]

Reference Book

11 June 2014

Region Features

421

39.2 Layer-Related
Region Features > Layer-Related
Layer-related region features evaluate the rst and second statistical moment (mean, standard deviation) of a regions pixel value.

39.2.1 Mean
Region Features > Layer Related > Mean
Mean image layer value of a given image layer within a region.
Editable Parameters

Region
Image layer
Parameters

ck is the mean intensity of image layer k

Expression

ck (R)
Feature Value Range
max
[cmin
R , cR ]

39.2.2 Standard Deviation

Region Features > Layer Related > Std Dev
The layer standard deviation of a given image layer within a region.
Editable Parameters

Region
Image layer
Parameters

is the standard deviation of image layer k

Reference Book

11 June 2014

Region Features

422

Expression

k (R)
Feature Value Range
max
[cmin
R , cR ]

39.3 Class-Related
Region Features > Class Related
Class-related region features provide information on all image objects of a given class
per region.

39.3.1 Area of Classified Objects

Region Features > Class Related > Area of Classified Objects
The absolute area in pixels, voxels, voxel series of all image objects classied as a given
class on a given image object level within a region.
Editable Parameters

Region
Class
Image object level
Parameters

#P is the total number of pixels

i is the the given image object level
R is the the given region
m is the the given class

Expression

#Pi (R, m)
Feature Value Range

[0, Rx Ry Rz Rt ]

Reference Book

11 June 2014

Region Features

423

39.3.2 Relative Area of Classified Objects

Region Features > Region-Related > Rel. Area of classified objects
The relative area in pixel, voxel, or voxel series of all image objects of a given class on a
given image object level within a region.
Editable Parameters

Region
Class
Image object Level
Parameters

#P is the total number of pixels

i is the the given image object level
R is the the given region
m is the the given class

Expression

#Pi (R, m)
#Pi (R)

Figure 39.1. Schematic display of the relative area of a class within a two dimensional region.

In this 55 sized region the feature values are:

Class Area of classified objects Relative area of classified objects

11 pixels

11/25 = 0.44

10 pixels

10/25 = 0.4

4 pixels

4/25 = 0.16

Feature Value Range

[0, 1]

Reference Book

11 June 2014

40 Image Registration Features

Region Features > Image Registration Features

40.1 Object-Related
Region Features > Image Registration Features > Object Related

40.1.1 Object Landmarks on the Map

Region Features > Image Registration Features > Object Related > Object Landmarks
on the Map
Returns the number of landmarks within the image object in the current map.

40.2 Scene-Related
Region Features > Image Registration Features > Scene Related

40.2.1 Landmarks on the Map

Region Features > Image Registration Features > Object Related > Landmarks on the
Map
Returns the number of landmarks with the current map.

424

41 Metadata
Metadata items can be used as a feature in rule set development. To do so, you have to
provide external metadata in the feature tree. If you are not using data import procedures
to convert external source metadata to internal metadata denitions, you can create individual features from a single metadata item. See also Object features > Object Metadata.

41.1 Active Slice Metadata item

Metadata > [Active Slice Metadata item]
For projects made up of multiple image slices, there may be different metadata values
on every slice. The Active Slice Metadata Item feature allows you to capture metadata
values for the image slice in the current active view. You can dene a new metadata item
by clicking on Create new Active Slice Metadata item.

41.1.1 Editable Parameters

41.2 Image Layer Metadata item

Metadata > [Image Layer Metadata item]
For projects with different metadata values on every image layer this feature allows you
to capture metadata values individually for each image layer in the current active view.
You can dene a new metadata item by clicking on Create new Image Layer Metadata
item.

41.2.1 Editable Parameters

Metadata Name: Name of the metadata source as used in the Modify Project dialog box. This name will be used in the feature tree for grouping corresponding

425

Metadata

426

metadata items. You can type a name of a not existing metadata source, to create a
feature group in advance.
Name: Name of the metadata item as used in the source data.
Type: Select the type of the metadata item: string, double, or integer.

41.3 Metadata Item

Metadata > [Metadata item]
If existing, all metadata items are listed in the feature tree. You can dene a new metadata
item by clicking on Create new Metadata item.

41.3.1 Editable Parameters

Map
Metadata Name: Name of the metadata source as used in the Modify Project dialog box. This name will be used in the feature tree for grouping corresponding
metadata items. You can type a name of a not existing metadata source, to create a
feature group in advance.
Metadata Access: Select the mode of accessing the metadata:
Direct: Metadata is accessed by the name of the metadata item
Conversion: Metadata is converted as dened in the Metadata Conversion
dialog box
Name: Name of the metadata item as used in the source data.
Value: If Metadata Access is set to Conversion , you can dene an internal metadata
item value , which can be different from the external value.
Type: Select the type of the metadata item: string, double, or integer.

Reference Book

11 June 2014

42 Feature Variables
Feature Variables have features as their values. Once a feature is assigned to a feature
variable, the feature variable can be used like that feature. It returns the same value as the
feature to which it points. It uses the unit of whatever feature is assigned as a variable.
It is possible to create a feature variable without a feature assigned, but the calculation
value would be invalid.
In a rule set, feature variables can be used like the corresponding feature.

42.1 Feature Variable

Feature Variables > [Feature variable]
If existing, all variables that use features as values are listed in the feature tree. You can
dene a new feature variable by clicking on Create new Feature Variable.

42.1.1 Editable Parameters

Name: Insert a name for the feature variable.
Value: Set the value of the feature variable.

427

43 Widget Parameters for

Architect Action Libraries
Listed below are all available widgets and parameters for use in Action Libraries. eCognition Developer 9.0 now has additional features that allow you to display TIFF images: if
you use these functions, ensure your images are stored in the same folder as your Action
Library, to ensure they will be visible to other users.

43.1 Add Checkbox

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Value Checked: The value of the variable if the checkbox is checked
Value Unchecked: The value of the variable if the checkbox is unchecked
Process on Change: The name of the process that will be executed when a user
changes a value

43.2 Add Drop-down List

428

Widget Parameters for Architect Action Libraries

429

Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Process on Selection Change: The name of the process that will be executed when
the selection changes
Items: Add or edit the items in the drop-down list

43.3 Add Button

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Process on Press: Enter the name of the process executed when the button is
pressed
Process on Release: Enter the name of the process executed when the button is
released
Button Text: Enter the text to be displayed on the button
Image: Allows you to display a TIFF image (transparency is supported)
Radio Button: If set to yes, a radio button is displayed
Ruleset: Navigate to the ruleset (.dcp le) containing the processes to be executed

43.4 Add Radio Button Row

Rule Set: Navigate to the ruleset (.dcp le) containing the processes to be executed
Radio Mode: If set to true, the buttons become radio buttons; if set to false they
are normal buttons
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Action Buttons:
Text: Enter the text to be displayed on the action button
Description: Enter a description to appear at the bottom of the description
pane. The text will also display as a tooltip. You may have a text description
or an image description, but not both

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

430

Description Image: Add a TIFF image to the description area (transparency

is supported). You may have a text description or an image description, but
not both
Process Path: Enter the process path to be executed by the action button
Process Path on Release: Enter the process path to be executed when the
button is released
Execute on a Separate Thread: If set to yes, the process will be executed on
a separate thread
Image File: Navigate to the path of the image le for the button image
Highlight Image File: Navigate to the path of the image le for the highlighted button image
Hot Key: Lets you dene a single hot key to execute an action
Radio Mode: Select push button to use this style specically; if from widget is selected, the radio mode from the widget is used

43.5 Add Toolbar

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Process on Change: The name of the process that will be executed when a user
changes a value

43.6 Add Editbox

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

431

Process on Change: The name of the process that will be executed when a user
changes a value

43.7 Add Editbox With Slider

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Process on Change: The name of the process that will be executed when a user
changes a value
Maximum Value: Enter the maximum value for the slider
Minumum Value: Enter the minimum value for the slider
Tick Frequency: Enter a value to dene how often tick marks appear next to the
slider
Jump Value: Enter a value to dene the increments when the slider is moved
Ruleset: Navigate to the ruleset (.dcp le) containing the processes to be executed

43.8 Add Select Class

Text: Enter the text that will be displayed to the user
Variable: Select or enter the variable that gets updated by this control
Available Classes: Select the classes to be made available (if none are selected,
classes are displayed based on action dependencies)
Process on Selection Change: Enter the name of the process to be executed when
a selection changes
Dependency Handling: The dependency effect of the selected class. Choose from
one of the following:
None
Required. This activates the parameter Dependency Error Message, which is
displayed when a dependency conict occurs. Use the tag #class within the
error text to identify the class name
Forbidden. This activates the parameter Dependency Error Message, which
is displayed when a dependency conict occurs. Use the tag #class within
the error text to identify the class name
Added
Removed
Allow Create New Class: Select yes to allow the user to create a new class from
the widget.

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

432

Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled

43.9 Add Select Feature

Text: Enter the text that will be displayed to the user
Variable: Select or enter the variable that gets updated by this control
Type: Select the lter type for selectable features all, scene features or object
features
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Process on Change: The name of the process that will be executed when a user
changes a value

43.10 Add Select Multiple Features

Text: Enter the text that will be displayed to the user
Image File: Navigate to the path of the image le for the button image
Highlight Image File: Navigate to the path of the image le for the highlighted
button image
Variable: Select or enter the variable that gets updated by this control
Type: Select the lter type to be used:
All
Object Features
Scene Features
Custom Filter this option brings up the Custom Feature Filter parameter,
which allows you to select which features are available for selection
Show Template Features: Select yes to show template features, while selecting
features

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

433

43.11 Add Select File

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable for File: Add or select the variable that gets updated by this control
File Filter: Add a lter for le type, for example *.tif or *.*
Type: Select the le to be selected:
Image
Thematic: If this option is selected, two more parameters appear:
* Variable for Attribute File, to be updated by this control
* Variable for Attribute ID Column, to be updated by this control
Mode: Select between open and save
Process on Change: The name of the process that will be executed when a user
changes a value

43.12 Add Select Level

Text: Enter the text that will be displayed to the user
Variable: Select or enter the variable that gets updated by this control
Show all Levels: If set to yes, action dependencies for listing available levels are
ignored
Process on Selection Change: The name of the process that will be executed when
the selection changes
Show Pixel Level: Select yes to show the pixel level

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

434

Dependency Handling: The dependency effect of the selected class. Choose from
one of the following:
None
Required. This activates the parameter Dependency Error Message, which is
displayed when a dependency conict occurs. Use the tag #class within the
error text to identify the class name
Forbidden. This activates the parameter Dependency Error Message, which
is displayed when a dependency conict occurs. Use the tag #class within
the error text to identify the class name
Added
Removed
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled

43.13 Add Select Image Layer

Text: Enter the text that will be displayed to the user
Variable: Select or enter the variable that gets updated by this control
Process on Selection Change: The name of the process that will be executed when
the selection changes
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled

43.14 Add Select Thematic Layer

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

435

43.15 Add Select Folder

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Process on Change: The name of the process that will be executed when a user
changes a value

43.16 Add Slider

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Variable: Select or enter the variable that gets updated by this control
Process on Change: The name of the process that will be executed when a user
changes a value
Maximum Value: Enter the maximum value for the slider
Minumum Value: Enter the minimum value for the slider

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

436

Tick Frequency: Enter a value to dene how often tick marks appear next to the
slider
Jump Value: Enter a value to dene the increments when the slider is moved
Ruleset: Navigate to the ruleset (.dcp le) containing the processes to be executed

43.17 Add Edit Layer Names

Text: Enter the text that will be displayed to the user

Tooltip: Enter the text to appear as a tooltip
Image File: Navigate to the path of the image le for the button image
Highlight Image File: Navigate to the path of the image le for the highlighted
button image
Button Size: Button size in pixels
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled

43.18 Add Layer Drop-down List

Text: Enter the text that will be displayed to the user
Description: Enter a description of your widget. The text will appear in the Description pane when the cursor hovers over the widget. You may have a text description or an image description, but not both
Description Image: Allows you to display a TIFF image in the Description pane
(transparency is supported). You may have a text description or an image description, but not both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Items: Add or edit the items in the drop-down list
Process on Selection Change: The name of the process that will be executed when
the selection changes

43.19 Add Manual Classification Buttons

Rule Set: Navigate to the ruleset (.dcp le) containing the processes to be executed
Mode: Choose between push button or checkbox

Reference Book

11 June 2014

Widget Parameters for Architect Action Libraries

437

Show Edit Classes Button: Select yes to display the Edit Classes button, which
allows users to change the class names and colors of widget classes
Image File for Edit Classes: Navigate to the path of the image le for the button
image
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled or disabled. A zero value denes the widget state as
disabled
Action Buttons:
Class: Assign a class to the button
Description: Enter a description to appear at the bottom of the description
pane. The text will also display as a tooltip
Description Image: Add a TIFF image to the description area (transparency
is supported). Uploading an image will replace the text in the description in
the description area. You may have a text description or an image description,
but not both
Process Path: Enter the process path to be executed by the action button
Process Path on Release: Enter the process path to be executed when the
button is released
Hot Key: Lets you dene a single hot key to execute an action

43.20 Add Select Array Items

Text: Enter the text that will be displayed to the user
Items: Enter a name then select the array that denes the selectable items
Selected Items: Enter a name then select the array that receives the selectable
items
Process on Change: The name of the process that will be executed when a user
changes a value
Description: Enter a description to appear at the bottom of the description pane.
The text will also display as a tooltip
Description Image: Add a TIFF image to the description area (transparency is
supported). Uploading an image will replace the text in the description in the description area. You may have a text description or an image description, but not
both
Show/Hide Variable: Enter or select the name of the variable that denes whether
the widget is visible. If you enter 0, the widget is hidden
Enable/Disable Variable: Enter or select the name of the variable that denes
whether the widget is enabled.

Reference Book

11 June 2014

44 General Reference
44.1 Use Variables as Features
The following variables can be used as features:

Scene variables
Object variables
Class variables
Feature variables.

They are displayed in the feature tree of, for example, the Feature View window or the
Select Displayed Features dialog box.

44.2 About Metadata as a Source of Information

Many image data formats include metadata or come with separate metadata les, which
provide information about the related image (for example, the acquisition time). This
metadata information can be converted into features, to use in image analysis.
The available metadata depends on the image reader or camera used, the industry-specic
environment, and the settings. Some industry-specic examples are:
Satellite image data containing metadata providing information on cloudiness
Microscopy image data that contains metadata on the magnication used. The
metadata provided can be displayed in the Image Object Information window and
can also be displayed in the feature tree, in the Feature View window or the Select
Displayed Features dialog box.

44.2.1 Convert Metadata and Add it to the Feature Tree

When importing data, you can load metadata source les to make them available within
the map. To provide external metadata items as features in the feature tree, you must
convert external source metadata to an internal metadata denition.
When developing rule sets, metadata denitions will be included in rule sets allowing
the serialization of metadata usage. Metadata conversion is available through import
functions and creating metadata item features.

438

General Reference

439

44.3 General Reference

44.3.1 Rendering a Displayed Image
1. The rst step reads out the displayed area from the selected image layers according
to the screen size and zoom settings. Then image layer equalization is applied. The
result is an 8-bit raw gray value image for each image layer. These gray value images are mixed into one raw RGB image by a layer mixer according to the current
layer mixing settings
2. Finally the image equalizing is applied to create the output RGB image that is
displayed on the screen.

Figure 44.1. The rendering process of an image displayed in the map view

Image Layer Equalization

Image layer equalization is part of the rendering process of the scene display within the
map view.
Image layer equalization maps the input data of an image layer which may have different
intensity ranges to the unied intensity range [0. . . 255] of an 8-bit gray value image.
For 8-bit data no image layer equalization is necessary. All other data types have to be
converted into an 8-bit representation at this step of the rendering process.
This function is implemented as a mapping of the input range to the display range of
[0. . . 255]. Image layer equalization can be either linear or manual. By default, the input
data is mapped to the gray value range by a linear function.
Data Type

Input Range

Mapping Function

cs = ck (no transformation)
cs = 255 ck
16-bit unsigned; 32-bit unsigned [0 . . . max2 (ck )]
max2 (ck )
cs = 255 (ck min2 (ck ))
16-bit signed; 32-bit signed
[min2 (ck ) . . . max2 (ck )]
max2 (ck ) min2 (ck )
cs = 255 (ck min10 (ck ))
32 bit oat
[min10 (ck ) . . . max10 (ck )]
max10 (ck ) min10 (ck )
8-bit

[0 . . . 255]

ck is the intensity value in image layer k

cs is the intensity value on the screen

Reference Book

11 June 2014

General Reference

440

min2 (ck ) = max{x : x = 2n ; x ck min } is the highest integer number that is a

power of 2 and darker than the darkest actual intensity value of all pixel values of
the selected image layer
max2 (ck ) = max{x : x = 2n ; x ck max } is the lowest integer number that is a power
of 2 and is brighter than the brightest actual intensity value of all pixel values of
the selected image layer
min10 (ck ) = max{x : x = 10n ; x ck min } is the highest integer number that is a
power of 10 and darker than the darkest actual intensity value of all pixel values of
the selected image layer
max10 (ck ) = max{x : x = 10n ; x ck max } is the lowest integer number that is a
power of 10 and is brighter than the brightest actual intensity value of all pixel
values of the selected image layer.
With manual image layer equalization you can specify the mapping function for each layer individually by dening both the input range
[cmin . . . cmax ] and the mapping function.
Manual Image Layer Equalization

Data Type

Mapping Function

Linear

cs = 255 (ck cmin ) (cmax cmin )

Linear inverse
Gamma correction (positive)
Gamma correction (negative)
Gamma correction (positive) inverse

cs = 255 255 (ck cmin ) (cmax cmin )

cs = 255 ((ck cmin ) (cmax cmin ))2

cs = 255 ((ck cmin ) (cmax cmin ))0.5

cs = 255 255 ((ck cmin ) (cmax cmin ))2

Gamma correction (negative) inverse cs = 255 255 ((ck cmin ) (cmax cmin ))0.5

ck is the intensity value in image layer k

cs is the intensity value on the screen
cmin is the smallest input intensity value (adjustable)
cmax is the largest input intensity value (adjustable).

Image Equalization

Image equalization is performed after all image layers are mixed into a raw RGB (red,
green, blue) image. If more than one image layer is assigned to one screen color (red,
green or blue) this approach leads to higher quality results. Where only one image layer
is assigned to each color, as is common, this approach is the same as applying equalization
to the individual raw layer gray value images
There are different modes for image equalization available. All of them rely on image
statistics. These are computed on the basis of a 256 256 pixel sized thumbnail of the
current raw RGB image.
No (None) equalization allows you to see the image data as it is, which can be
helpful at the beginning of rule set development, when looking for an approach. The
output from the image layer mixing is displayed without further modication.
None

[0 . . . 255] is the input range

Reference Book

11 June 2014

General Reference

441

[0 . . . 255] is the mapping function

Linear equalization with 1.00% is the default for new scenes. Commonly it displays images with a higher contrast as without image equalization.
Linear Equalization

Linear equalization maps each color red, green, and blue from an input range
[cmin . . . cmax ] to the available screen intensity range [0 . . . 255] by a linear mapping. The
input range can be modied by the percentage parameter p. The input range is computed
such that p per cent of the pixels are not part of the input range. In case p = 0 this means
the range of used color values is stretched to the range [0 . . . 255] . For p > 0 the mapping
ignores 2p per cent of the darkest pixels and 2p per cent of the brightest pixels. In many
cases a small value of p leads to better results because the available color range can be
better used for the relevant data by ignoring the outliers.
max{c : #{(x, y) : ck (x, y) < cmin }
(sx sy) 2p
min{c : #{(x, y) : ck (x, y) > cmax }
cmax =
(sx sy) 2p
[cmin . . . cmax] is the input
range
ck cmin
cs = 255
is the mapping function.
cmax cmin
cmin =

With its default parameter of 3.0, standard deviation renders a similar display as linear equalization. Use a parameter around 1.0 for an exclusion
of dark and bright outliers.
Standard Deviation Equalization

Standard deviation equalization maps the input range to the available screen intensity
range [0 . . . 255] by a linear mapping. The input range [cmin . . . cmax ] can be modied by
the width p. The input range is computed such that the center of the input range represents
the mean value of the pixel intensities mean (ck ). The left and right borders of the input
range are computed by taking n times the standard deviation k to the left and the right.
You can modify the parameter n.
cmin = mean(ck ) n k
cmax = mean(ck ) + n k
[cmin . . . cmax] is the input
range
ck cmin
is the mapping function.
cs = 255
cmax cmin
Gamma correction equalization is used to improve the
contrast of dark or bright areas by spreading the corresponding gray values. Gamma correction equalization maps the input range to the available screen intensity range [0 . . . 255]
by a polynomial mapping. The input range [cmin . . . cmax ] cannot be be modied and is
dened by the smallest and the largest existing pixel values.
Gamma Correction Equalization

min

cmin = c k the smallest existing pixel values

max
cmax = c k the largest existing pixel values
[cmin . . . cmax] is the input
range
ck cmin n
is the mapping function.
cs = 255
cmax cmin
Reference Book

11 June 2014

General Reference

442

You can be modify the exponent of the mapping function by editing the equalization
parameter. Values of n less than 1 emphasize darker regions of the image, values larger
than one emphasize darker areas of the image. A value from n = 1 represents the linear
case.
Histogram equalization is well suited for LANDSAT images but
can lead to substantial over-stretching on many normal images. It can be helpful in cases
you want to display dark areas with more contrast. Histogram equalization maps the input
range to the available screen intensity range [0 . . . 255] by a nonlinear function. Simply
said, the mapping is dened by the property that each color value of the output image
represents the same number of pixels. The respective algorithm is more complex and can
be found in standard image processing literature.
Histogram Equalization

Manual image layer equalization allows you to control

equalization in detail. For each image layer individually, you can set the equalization
method specifying the mapping function. Further you can dene the input range by setting minimum and maximum values.
Manual Image Layer Equalization

Reference Book

11 June 2014

Acknowledgments
Portions of this product are based in part on the third-party software components. Trimble
is required to include the following text, with software and distributions.

The Visualization Toolkit (VTK) Copyright

This is an open-source copyright as follows:
Copyright 19932006 Ken Martin, Will Schroeder and Bill Lorensen.
All rights reserved.
Redistribution and use in source and binary forms, with or without modication, are
permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of
conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list
of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither name of Ken Martin, Will Schroeder, or Bill Lorensen nor the names of
any contributors may be used to endorse or promote products derived from this
software without specic prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

443

Acknowledgments

444

ITK Copyright
Copyright 19992003 Insight Software Consortium
All rights reserved.
Redistribution and use in source and binary forms, with or without modication, are
permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of
conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list
of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of the Insight Software Consortium nor the names of its contributors may be used to endorse or promote products derived from this software
without specic prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

python/tests/test_doctests.py
Copyright 2007, Sean C. Gillies, [email protected]
All rights reserved.
Redistribution and use in source and binary forms, with or without modication, are
permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of
conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list
of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of Sean C. Gillies nor the names of its contributors may be used
to endorse or promote products derived from this software without specic prior
written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUD-

Reference Book

11 June 2014

Acknowledgments

445

ING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

src/Verson.rc
Copyright 2005, Frank Warmerdam, [email protected]
All rights reserved.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation les (the Software), to deal in the Software without
restriction, including without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom
the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or
substantial portions of the Software.
THE SOFTWARE IS PROVIDED AS IS, WITHOUT WARRANTY OF ANY
KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE
AND NON-INFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE
OR OTHER

src/gt_wkt_srs.cpp
Copyright 1999, Frank Warmerdam, [email protected]
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation les (the Software), to deal in the Software without
restriction, including without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom
the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or
substantial portions of the Software.
THE SOFTWARE IS PROVIDED AS IS, WITHOUT WARRANTY OF ANY
KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE

Reference Book

11 June 2014

eCognition Developer 9.0

446

AND NON-INFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE
OR OTHER DEALINGS IN THE SOFTWARE.

2014 Trimble Germany GmbH

Page collection published 11 June 2014 Typeset by Wikipublisher

Reference Book

11 June 2014

Delixi Inverter Manual
75% (4)
Delixi Inverter Manual
152 pages
QGIS 2.14 QGISTrainingManual en
100% (1)
QGIS 2.14 QGISTrainingManual en
639 pages
Vic 2D 7.0 Manual
No ratings yet
Vic 2D 7.0 Manual
168 pages
Program Enrollment Test Quiz - WorldQuant University
No ratings yet
Program Enrollment Test Quiz - WorldQuant University
4 pages
MPX200 Multifunction Router User Guide
No ratings yet
MPX200 Multifunction Router User Guide
251 pages
Anki Manual
No ratings yet
Anki Manual
195 pages
PNG Department of Works Standard Survey Instructions 1985
No ratings yet
PNG Department of Works Standard Survey Instructions 1985
75 pages
Rivers of India
No ratings yet
Rivers of India
37 pages
UserGuide PDF
No ratings yet
UserGuide PDF
270 pages
MIKEView User Guide
No ratings yet
MIKEView User Guide
159 pages
Non Convex Optimization PDF
No ratings yet
Non Convex Optimization PDF
204 pages
Opencv2refman PDF
No ratings yet
Opencv2refman PDF
899 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
913 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
929 pages
Opencv Reference Manual PDF
No ratings yet
Opencv Reference Manual PDF
817 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
819 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
931 pages
QGIS 3.10 UserGuide Fi
No ratings yet
QGIS 3.10 UserGuide Fi
1,165 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
929 pages
12d Advanced Storm Water
No ratings yet
12d Advanced Storm Water
72 pages
Manual de Referencia Opencv
No ratings yet
Manual de Referencia Opencv
915 pages
Colab Tutorial
No ratings yet
Colab Tutorial
21 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
553 pages
Enhancement 18 01 12
No ratings yet
Enhancement 18 01 12
33 pages
TextMate Ruby/Rails Cheat Sheet
100% (2)
TextMate Ruby/Rails Cheat Sheet
3 pages
WT Lab Manual
100% (1)
WT Lab Manual
44 pages
Algebra FX 2.0 Plus FX 1.0 Plus: User's Guide
No ratings yet
Algebra FX 2.0 Plus FX 1.0 Plus: User's Guide
455 pages
David Tarboton, Dan Watson, Rob Wallace: Dtarb@usu - Edu
No ratings yet
David Tarboton, Dan Watson, Rob Wallace: Dtarb@usu - Edu
49 pages
WSUD Technical Design Guidelines Online PDF
No ratings yet
WSUD Technical Design Guidelines Online PDF
351 pages
Principles of Remote Sensing
No ratings yet
Principles of Remote Sensing
410 pages
UserGuide PDF
No ratings yet
UserGuide PDF
256 pages
BITH203 Computer Repair and Maintenance
No ratings yet
BITH203 Computer Repair and Maintenance
191 pages
SMPTE ST 291-12011 Ancillary-Data-Packet-And-Space-Formatting
100% (1)
SMPTE ST 291-12011 Ancillary-Data-Packet-And-Space-Formatting
17 pages
Nepal Engg Council Act, 2055
No ratings yet
Nepal Engg Council Act, 2055
22 pages
Proposed PHD in Data Science
No ratings yet
Proposed PHD in Data Science
166 pages
Opencv2refman CPP
No ratings yet
Opencv2refman CPP
409 pages
Logical Reasoning-Introduction
No ratings yet
Logical Reasoning-Introduction
28 pages
Avaya CMS Maintaining and Troubleshooting 19.2 March 2021
No ratings yet
Avaya CMS Maintaining and Troubleshooting 19.2 March 2021
308 pages
Geoprocessing
No ratings yet
Geoprocessing
6 pages
STEEL1
No ratings yet
STEEL1
2 pages
CV2019
No ratings yet
CV2019
152 pages
2018學位生各院教授名單List of Professors for Degree-seeking Students-201804updated
No ratings yet
2018學位生各院教授名單List of Professors for Degree-seeking Students-201804updated
13 pages
Department Wise - SMU 2.7.1
No ratings yet
Department Wise - SMU 2.7.1
95 pages
CH 2 Multiple Regression S
No ratings yet
CH 2 Multiple Regression S
78 pages
Relief Features and Contour Lines
No ratings yet
Relief Features and Contour Lines
6 pages
Porosity: Ground - Water Hydraulic M. Sc. Shahad Abdulkareem Raheem
No ratings yet
Porosity: Ground - Water Hydraulic M. Sc. Shahad Abdulkareem Raheem
11 pages
First Order Systems
No ratings yet
First Order Systems
40 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
643 pages
Lecture 15
No ratings yet
Lecture 15
37 pages
Advanced Fluid Mechanics: Lecture Notes Module-IV
No ratings yet
Advanced Fluid Mechanics: Lecture Notes Module-IV
20 pages
A Survey On Sentiment Analysis Methods Applications and Challenges
No ratings yet
A Survey On Sentiment Analysis Methods Applications and Challenges
50 pages
Opencv 2 Refman
No ratings yet
Opencv 2 Refman
553 pages
Final Intern Apoorva
No ratings yet
Final Intern Apoorva
38 pages
Practical Research Paper Smart Dustbin 12 Euler 1
No ratings yet
Practical Research Paper Smart Dustbin 12 Euler 1
41 pages
CD GTU Study Material Presentations Unit-1 27062020072512AM
No ratings yet
CD GTU Study Material Presentations Unit-1 27062020072512AM
41 pages
Object Oriented Programming in ANSI C by Balaguruswamy PDF Download
No ratings yet
Object Oriented Programming in ANSI C by Balaguruswamy PDF Download
29 pages
Chapter 4 2023 2
No ratings yet
Chapter 4 2023 2
25 pages
Flood Model Study of Wardha River MJB
No ratings yet
Flood Model Study of Wardha River MJB
19 pages
As Cfe Interop 61850 en PDF
No ratings yet
As Cfe Interop 61850 en PDF
29 pages
Introduction To Gis Using Arcgis: Student Edition
No ratings yet
Introduction To Gis Using Arcgis: Student Edition
6 pages
Service Based Interface Development in 5G CN: 5 OAI Workshop
No ratings yet
Service Based Interface Development in 5G CN: 5 OAI Workshop
23 pages
18AI71 - AAI INTERNAL 2 QB Answers
No ratings yet
18AI71 - AAI INTERNAL 2 QB Answers
16 pages
Powerpoint Template: " Add Your Company Slogan "
No ratings yet
Powerpoint Template: " Add Your Company Slogan "
20 pages
Fuchs Overdrive Supreme 50 Manual
No ratings yet
Fuchs Overdrive Supreme 50 Manual
19 pages
Map Basic User Guide
No ratings yet
Map Basic User Guide
319 pages
A Contribution To New ALE 2D Method Validation - Ansys
No ratings yet
A Contribution To New ALE 2D Method Validation - Ansys
12 pages
NUS MA4268 Ch1
No ratings yet
NUS MA4268 Ch1
9 pages
OB1 - : Name: Family: Author: Block Version: Time Stamp Code: Interface: Lengths (Block/logic/data)
No ratings yet
OB1 - : Name: Family: Author: Block Version: Time Stamp Code: Interface: Lengths (Block/logic/data)
11 pages
Demo
No ratings yet
Demo
9 pages
Exit Card l9 1st Year
No ratings yet
Exit Card l9 1st Year
2 pages
Discover Qgis3 Toc
No ratings yet
Discover Qgis3 Toc
7 pages
Process List
No ratings yet
Process List
14 pages
Zamfira Ioana Ruxandra - Raport
No ratings yet
Zamfira Ioana Ruxandra - Raport
10 pages
Effects of Social Media On Grade 11 Students of Voctech
No ratings yet
Effects of Social Media On Grade 11 Students of Voctech
11 pages
S3400 PDF
No ratings yet
S3400 PDF
16 pages
Signals and Systems Question Bank
No ratings yet
Signals and Systems Question Bank
4 pages
Intelligent Public Transportation Systems: A Review of Architectures and Enabling Technologies
No ratings yet
Intelligent Public Transportation Systems: A Review of Architectures and Enabling Technologies
8 pages
Deep Learning Models For Plant Disease Detection and Diagnosis.
No ratings yet
Deep Learning Models For Plant Disease Detection and Diagnosis.
9 pages
The Hash Table Data Structure: Mugurel Ionuț Andreica Spring 2012
No ratings yet
The Hash Table Data Structure: Mugurel Ionuț Andreica Spring 2012
9 pages
Karthik CV 2025-Updated
No ratings yet
Karthik CV 2025-Updated
5 pages
Wolfsmilkie - Tumblr Blog Tumgik
No ratings yet
Wolfsmilkie - Tumblr Blog Tumgik
4 pages
Circular Diaphragm
No ratings yet
Circular Diaphragm
4 pages
Photogrammetry: Prepared by Mitsuru Nasu, PH.D
No ratings yet
Photogrammetry: Prepared by Mitsuru Nasu, PH.D
2 pages
Learning, Understanding and Interaction in Videos: Manmohan Chandraker
No ratings yet
Learning, Understanding and Interaction in Videos: Manmohan Chandraker
1 page
Image Analysis, Classification and Change Detection in Remote Sensing
No ratings yet
Image Analysis, Classification and Change Detection in Remote Sensing
6 pages
True, False, Not Given PDF
No ratings yet
True, False, Not Given PDF
2 pages
Microsoft Nav 2009 Part A
No ratings yet
Microsoft Nav 2009 Part A
3 pages
Od 427305939568238100
No ratings yet
Od 427305939568238100
1 page
Dragonpay Payment Instruction
No ratings yet
Dragonpay Payment Instruction
1 page
Advanced Multiplayer Game Development with Ureal Engine 5: A Comprehensive Guide to C++ Scripting
From Everand
Advanced Multiplayer Game Development with Ureal Engine 5: A Comprehensive Guide to C++ Scripting
Vladimir Kiselev
No ratings yet