Download the full version of the ebook at

https://ebookmass.com

3-D Computer Vision: Principles, Algorithms

and Applications 1st Edition Yu-Jin Zhang

https://ebookmass.com/product/3-d-computer-vision-
principles-algorithms-and-applications-1st-
edition-yu-jin-zhang/

Explore and download more ebook at https://ebookmass.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Computer vision: theory, algorithms, practicalities Fifth

Edition Davies

https://ebookmass.com/product/computer-vision-theory-algorithms-
practicalities-fifth-edition-davies/

testbankdeal.com

Hidden Semi-Markov models : theory, algorithms and

applications 1st Edition Yu

https://ebookmass.com/product/hidden-semi-markov-models-theory-
algorithms-and-applications-1st-edition-yu/

testbankdeal.com

New Optimization Algorithms and their Applications: Atom-

Based, Ecosystem-Based and Economics-Based 1st Edition
Zhang
https://ebookmass.com/product/new-optimization-algorithms-and-their-
applications-atom-based-ecosystem-based-and-economics-based-1st-
edition-zhang/
testbankdeal.com

Building Computer Vision Applications Using Artificial

Neural Networks, 2nd Edition Shamshad Ansari

https://ebookmass.com/product/building-computer-vision-applications-
using-artificial-neural-networks-2nd-edition-shamshad-ansari/

testbankdeal.com
Feature extraction and image processing for computer
vision Fourth Edition Aguado

https://ebookmass.com/product/feature-extraction-and-image-processing-
for-computer-vision-fourth-edition-aguado/

testbankdeal.com

Probabilistic graphical models for computer vision Ji Q

https://ebookmass.com/product/probabilistic-graphical-models-for-
computer-vision-ji-q/

testbankdeal.com

Transformers for Natural Language Processing and Computer

Vision, Third Edition Denis Rothman

https://ebookmass.com/product/transformers-for-natural-language-
processing-and-computer-vision-third-edition-denis-rothman/

testbankdeal.com

Computer Vision Technology for Food Quality Evaluation,

Second Edition Sun

https://ebookmass.com/product/computer-vision-technology-for-food-
quality-evaluation-second-edition-sun/

testbankdeal.com

Algorithms For Dummies (For Dummies (Computer/Tech)) 2nd

Edition Mueller

https://ebookmass.com/product/algorithms-for-dummies-for-dummies-
computer-tech-2nd-edition-mueller/

testbankdeal.com
Yu-Jin Zhang

3-D Computer
Vision
Principles, Algorithms and Applications
3-D Computer Vision
Yu-Jin Zhang

3-D Computer Vision

Principles, Algorithms and Applications
Yu-Jin Zhang
Department of Electronic Engineering
Tsinghua University
Beijing, China

ISBN 978-981-19-7579-0 ISBN 978-981-19-7580-6 (eBook)

https://doi.org/10.1007/978-981-19-7580-6

Jointly published with Publishing House of Electronics Industry

The print edition is not for sale in China (Mainland). Customers from China (Mainland) please order the
print book from: Publishing House of Electronics Industry.

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore
Pte Ltd. 2023

Jointly published with Publishing House of Electronics Industry, Beijing, China

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publishers, the authors, and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publishers nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publishers remain neutral with
regard to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

This book is a specialized textbook that introduces the basic principles, typical
methods, and practical techniques of 3-D computer vision. It can provide an
advanced/second course service of computer vision for graduate students of related
majors in higher engineering colleges and universities, after studying 2D Computer
Vision: Principles, Algorithms and Applications.
This book mainly covers the higher level of computer vision from the selection of
materials. This book is self-contained, mainly for information majors, but also takes
into account learners of different professional backgrounds, and also considers the
needs of self-study readers. After learning the fundamental content of computer
vision in this book, readers can carry out scientific research and solve more and even
difficult specific problems in practical applications.
This book pays more attention to practicality in writing. Considering that com-
puter vision technology has been involved in many professional fields in recent
years, but many working people are not specialized in computer vision technology,
this book does not emphasize the theoretical system too much, minimizes the
formula derivation, and focuses on commonly used techniques. This book has
many sample questions and uses intuitive explanation to help readers understand
abstract concepts. A subject index list is given at the end of the book, and those
corresponding terms are marked in bold in the text.
This book provides a large number of self-test questions (including hints and
answers). In terms of purpose: on the one hand, it is convenient for self-study to
judge whether they have mastered the key content; on the other hand, it is also
convenient for teachers to carry out online teaching and strengthen teacher-student
interaction during lectures. The types of questions are multiple-choice questions,
which can be easily judged by a computer. In terms of content, many questions
express the basic concepts in a different way, supplementing the text, so that learners
can deepen their understanding. Some questions list descriptions that are similar but
not the same, or even have opposite meanings. Through dialectical thinking of pros
and cons, learners can also deeply understand the essence. The hints have been
provided for all self-test questions, allowing readers to obtain more information to

v
vi Preface

further identify the meaning of the questions. At the same time, each question can be
said to be divided into two levels in this way. Readers can complete the self-test after
reading the hints to show that they basically understand it, and to complete the self-
test without looking at the hints indicates that they have an even better grasp.
From the structure of this book, there are 12 chapters in total, plus one appendix,
answers to self-test questions, and subject index. Under these 15 first-level headings,
there are a total of 103 second-level headings (sections), and there are 141 third-level
headings (subsections) underneath. The book has a total of about 500,000 words
(including pictures, drawings, tables, formulas, etc.) and a total of (numbered)
228 figures, 22 tables, and 566 formulas. In order to facilitate teaching and learning,
this book provides a total of 68 examples of various types and 157 self-test questions
(all with hints and answers). In addition, there are a list of more than 100 directly
related references and a list of more than 500 subject terms for indexing, at the end of
the book.
This book can consider three aspects from the knowledge requirements of the
prerequisite courses: (1) Mathematics: including linear algebra and matrix theory, as
well as basic knowledge about statistics, probability theory, and random modeling;
(2) Computer science: including the mastery of computer software technology, the
understanding of computer structure system, and the application of computer pro-
gramming methods; (3) Electronics: On the one hand, the characteristics and prin-
ciples of electronic equipment; on the other hand, circuit design and other content. In
addition, the book 2D Computer Vision: Principles, Algorithms and Applications
can be counted as the discipline prerequisite of this book.
Thanks to the editors of the publisher for carefully composing the manuscript,
seriously reviewing, and attentively modifying.
Finally, the author thanks his wife Yun HE and daughter Heming ZHANG for
their understanding and support in all aspects.

Beijing, China Yu-Jin Zhang

Contents

1 Computer Vision Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Human Vision and Characteristics . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Visual Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.2 Brightness Properties of Vision . . . . . . . . . . . . . . . . . . 3
1.1.3 Spatial Properties of Vision . . . . . . . . . . . . . . . . . . . . . 6
1.1.4 Temporal Properties of Vision . . . . . . . . . . . . . . . . . . . 8
1.1.5 Visual Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 Computer Vision Theory and Framework . . . . . . . . . . . . . . . . . 13
1.2.1 Reaserch Goals, Tasks, and Methods of Computer
Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.2 Visual Computational Theory . . . . . . . . . . . . . . . . . . . 14
1.2.3 Framework Problems and Improvements . . . . . . . . . . . 20
1.3 Three-Dimensional Vision System and Image Technology . . . . . 23
1.3.1 Three-Dimensional Vision System Process . . . . . . . . . . 23
1.3.2 Computer Vision and Image Technology Levels . . . . . . 24
1.3.3 Image Technology Category . . . . . . . . . . . . . . . . . . . . 25
1.4 Overview of the Structure and Content of This Book . . . . . . . . . 27
1.4.1 Structural Framework and Content of This Book . . . . . 27
1.4.2 Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 29
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2 Camera Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.1 Linear Camera Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.1.1 Complete Imaging Model . . . . . . . . . . . . . . . . . . . . . . 38
2.1.2 Basic Calibration Procedure . . . . . . . . . . . . . . . . . . . . 40
2.1.3 Internal and External Parameters . . . . . . . . . . . . . . . . . 41
2.2 Non-Linear Camera Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2.1 Type of Distortion . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

vii
viii Contents

2.2.2 Calibration Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

2.2.3 Classification of Calibration Methods . . . . . . . . . . . . . 48
2.3 Traditional Calibration Methods . . . . . . . . . . . . . . . . . . . . . . . . 49
2.3.1 Basic Steps and Parameters . . . . . . . . . . . . . . . . . . . . . 49
2.3.2 Two-Stage Calibration Method . . . . . . . . . . . . . . . . . . 51
2.3.3 Precision Improvement . . . . . . . . . . . . . . . . . . . . . . . . 56
2.4 Self-Calibration Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 61
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3 Three-Dimensional Image Acquisition . . . . . . . . . . . . . . . . . . . . . . . 67
3.1 High-Dimensional Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.2 Depth Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.2.1 Depth Image and Grayscale Image . . . . . . . . . . . . . . . 69
3.2.2 Intrinsic Image and Non-Intrinsic Image . . . . . . . . . . . 70
3.2.3 Depth Imaging Modes . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3 Direct Depth Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3.1 Time-of-Flight Method . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.2 Structured Light Method . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.3 Moiré Contour Stripes Method . . . . . . . . . . . . . . . . . . 79
3.3.4 Simultaneous Acquisition of Depth and Brightness
Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.4 Stereo Vision Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.1 Binocular Horizontal Mode . . . . . . . . . . . . . . . . . . . . . 85
3.4.2 Binocular Convergence Horizontal Mode . . . . . . . . . . . 89
3.4.3 Binocular Axial Mode . . . . . . . . . . . . . . . . . . . . . . . . 92
3.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 95
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4 Video Image and Motion Information . . . . . . . . . . . . . . . . . . . . . . . 101
4.1 Video Basic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.1.1 Video Expression and Model . . . . . . . . . . . . . . . . . . . 102
4.1.2 Video Display and Format . . . . . . . . . . . . . . . . . . . . . 104
4.1.3 Color TV System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2 Motion Classification and Representation . . . . . . . . . . . . . . . . . 108
4.2.1 Motion Classification . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.2.2 Motion Vector Field Representation . . . . . . . . . . . . . . 109
4.2.3 Motion Histogram Representation . . . . . . . . . . . . . . . . 111
4.2.4 Motion Track Description . . . . . . . . . . . . . . . . . . . . . . 113
4.3 Motion Information Detection . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.3.1 Motion Detection Based on Camera Model . . . . . . . . . 114
4.3.2 Frequency Domain Motion Detection . . . . . . . . . . . . . 118
4.3.3 Detection of Movement Direction . . . . . . . . . . . . . . . . 121
Contents ix

4.4 Motion-Based Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

4.4.1 Motion Detection-Based Filtering . . . . . . . . . . . . . . . . 123
4.4.2 Motion Compensation-Based Filtering . . . . . . . . . . . . . 124
4.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 130
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5 Moving Object Detection and Tracking . . . . . . . . . . . . . . . . . . . . . . 135
5.1 Differential Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.1.1 Calculation of Difference Image . . . . . . . . . . . . . . . . . 136
5.1.2 Calculation of Accumulative Difference Image . . . . . . . 138
5.2 Background Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.2.1 Basic Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.2.2 Typical Practical Methods . . . . . . . . . . . . . . . . . . . . . . 140
5.2.3 Effect Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.3 Optical Flow Field and Motion . . . . . . . . . . . . . . . . . . . . . . . . 145
5.3.1 Optical Flow Equation . . . . . . . . . . . . . . . . . . . . . . . . 145
5.3.2 Optical Flow Estimation with Least Square Method . . . 146
5.3.3 Optical Flow in Motion Analysis . . . . . . . . . . . . . . . . . 147
5.3.4 Dense Optical Flow Algorithm . . . . . . . . . . . . . . . . . . 150
5.4 Moving Object Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.4.1 Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
5.4.2 Particle Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
5.4.3 Mean Shift and Kernel Tracking . . . . . . . . . . . . . . . . . 160
5.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 164
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
6 Binocular Stereo Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.1 Stereo Vision Process and Modules . . . . . . . . . . . . . . . . . . . . . 170
6.1.1 Camera Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.1.2 Image Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.1.3 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.1.4 Stereo Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.1.5 3-D Information Recovery . . . . . . . . . . . . . . . . . . . . . 172
6.1.6 Post-Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.2 Region-Based Stereo Matching . . . . . . . . . . . . . . . . . . . . . . . . 173
6.2.1 Template Matching . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.2.2 Stereo Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
6.3 Feature-Based Stereo Matching . . . . . . . . . . . . . . . . . . . . . . . . 185
6.3.1 Basic Steps and Methods . . . . . . . . . . . . . . . . . . . . . . 186
6.3.2 Scale Invariant Feature Transformation . . . . . . . . . . . . 190
6.3.3 Dynamic Programming Matching . . . . . . . . . . . . . . . . 192
6.4 Error Detection and Correction of Parallax Map . . . . . . . . . . . . 194
6.4.1 Error Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
 Visit https://ebookmass.com
now to explore a rich
collection of eBooks and enjoy
exciting offers!
x Contents

6.4.2 Error Correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

6.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 198
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7 Monocular Multiple Image Recovery . . . . . . . . . . . . . . . . . . . . . . . 205
7.1 Photometric Stereo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
7.1.1 Light Source, Scenery, Lens . . . . . . . . . . . . . . . . . . . . 207
7.1.2 Scene Brightness and Image Brightness . . . . . . . . . . . . 208
7.1.3 Surface Reflection Characteristics and Brightness . . . . . 213
7.2 Shape from Illumination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
7.2.1 Representation of the Surface Orientation
of a Scene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
7.2.2 Reflectance Map and Brightness Constraint
Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
7.2.3 Solution of Photometric Stereo . . . . . . . . . . . . . . . . . . 222
7.3 Optical Flow Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
7.3.1 Optical Flow and Motion Field . . . . . . . . . . . . . . . . . . 226
7.3.2 Solving Optical Flow Equation . . . . . . . . . . . . . . . . . . 229
7.4 Shape from Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 240
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
8 Monocular Single Image Recovery . . . . . . . . . . . . . . . . . . . . . . . . . 247
8.1 Shape from Shading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
8.1.1 Shading and Orientation . . . . . . . . . . . . . . . . . . . . . . . 248
8.1.2 Gradient Space Method . . . . . . . . . . . . . . . . . . . . . . . 250
8.2 Solving Brightness Equation . . . . . . . . . . . . . . . . . . . . . . . . . . 253
8.2.1 Linearity Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
8.2.2 Rotational Symmetry Case . . . . . . . . . . . . . . . . . . . . . 256
8.2.3 The General Case of Smoothness Constraints . . . . . . . . 258
8.3 Shape from Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
8.3.1 Monocular Imaging and Distortion . . . . . . . . . . . . . . . 260
8.3.2 Orientation Restoration from the Change of Texture . . . 263
8.4 Detection of Texture Vanishing Points . . . . . . . . . . . . . . . . . . . 271
8.4.1 Detecting the Vanishing Point of Line Segment
Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
8.4.2 Determine the Vanishing Point Outside the Image . . . . 273
8.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 276
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
9 Three-Dimensional Scenery Representation . . . . . . . . . . . . . . . . . . 283
9.1 Local Features of the Surface . . . . . . . . . . . . . . . . . . . . . . . . . . 284
9.1.1 Surface Normal Section . . . . . . . . . . . . . . . . . . . . . . . 284
Contents xi

9.1.2 Surface Principal Curvature . . . . . . . . . . . . . . . . . . . . 285

9.1.3 Mean Curvature and Gaussian Curvature . . . . . . . . . . . 286
9.2 Three-Dimensional Surface Representation . . . . . . . . . . . . . . . . 288
9.2.1 Parameter Representation . . . . . . . . . . . . . . . . . . . . . . 288
9.2.2 Surface Orientation Representation . . . . . . . . . . . . . . . 291
9.3 Construction and Representation of Iso-surfaces . . . . . . . . . . . . 294
9.3.1 Marching Cube Algorithm . . . . . . . . . . . . . . . . . . . . . 295
9.3.2 Wrapper Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . 298
9.4 Interpolating Three-Dimensional Surfaces from Parallel
Contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
9.4.1 Contour Interpolation and Tiling . . . . . . . . . . . . . . . . . 300
9.4.2 Problems That May Be Encountered . . . . . . . . . . . . . . 302
9.4.3 Delaunay Triangulation and Neighborhood Voronoï
Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
9.5 Three-Dimensional Entity Representation . . . . . . . . . . . . . . . . . 306
9.5.1 Basic Representation Scheme . . . . . . . . . . . . . . . . . . . 306
9.5.2 Generalized Cylinder Representation . . . . . . . . . . . . . . 309
9.6 Key Points and References for Each Section . . . . . . . . . . . . . . . 311
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
10 Generalized Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
10.1 Matching Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
10.1.1 Matching Strategies and Categories . . . . . . . . . . . . . . . 318
10.1.2 Matching and Registration . . . . . . . . . . . . . . . . . . . . . 321
10.1.3 Matching Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 323
10.2 Object Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
10.2.1 Measure of Matching . . . . . . . . . . . . . . . . . . . . . . . . . 324
10.2.2 Corresponding Point Matching . . . . . . . . . . . . . . . . . . 327
10.2.3 String Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
10.2.4 Matching of Inertia Equivalent Ellipses . . . . . . . . . . . . 329
10.2.5 Shape Matrix Matching . . . . . . . . . . . . . . . . . . . . . . . 331
10.3 Dynamic Pattern Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
10.3.1 Matching Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
10.3.2 Absolute Pattern and Relative Pattern . . . . . . . . . . . . . 334
10.4 Graph Theory and Graph Matching . . . . . . . . . . . . . . . . . . . . . 336
10.4.1 Introduction to Graph Theory . . . . . . . . . . . . . . . . . . . 336
10.4.2 Graph Isomorphism and Matching . . . . . . . . . . . . . . . . 340
10.5 Line Drawing Signature and Matching . . . . . . . . . . . . . . . . . . . 343
10.5.1 Contour Marking . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
10.5.2 Structural Reasoning . . . . . . . . . . . . . . . . . . . . . . . . . 344
10.5.3 Labeling via Backtracking . . . . . . . . . . . . . . . . . . . . . 346
10.6 Key Points and References for Each Section . . . . . . . . . . . . . . . 347
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
xii Contents

11 Knowledge and Scene Interpretation . . . . . . . . . . . . . . . . . . . . . . . . 355

11.1 Scene Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
11.1.1 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
11.1.2 Attribute Hypergraph . . . . . . . . . . . . . . . . . . . . . . . . . 358
11.1.3 Knowledge-Based Modeling . . . . . . . . . . . . . . . . . . . . 360
11.2 Logic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
11.2.1 Predicate Calculation Rules . . . . . . . . . . . . . . . . . . . . . 362
11.2.2 Inference by Theorem Proving . . . . . . . . . . . . . . . . . . 366
11.3 Fuzzy Reasoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
11.3.1 Fuzzy Sets and Fuzzy Operations . . . . . . . . . . . . . . . . 370
11.3.2 Fuzzy Reasoning Method . . . . . . . . . . . . . . . . . . . . . . 372
11.4 Scene Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
11.4.1 Bag-of-Words/Feature Model . . . . . . . . . . . . . . . . . . . 375
11.4.2 pLSA Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
11.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 383
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
12 Spatial-Temporal Behavior Understanding . . . . . . . . . . . . . . . . . . . 389
12.1 Spatial-Temporal Technology . . . . . . . . . . . . . . . . . . . . . . . . . 390
12.1.1 New Research Field . . . . . . . . . . . . . . . . . . . . . . . . . . 390
12.1.2 Multiple Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
12.2 Spatial-Temporal Interest Point Detection . . . . . . . . . . . . . . . . . 392
12.2.1 Detection of Points of Interest in Space . . . . . . . . . . . . 392
12.2.2 Detection of Points of Interest in Space and Time . . . . . 393
12.3 Spatial-Temporal Dynamic Trajectory Learning and
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
12.3.1 Automatic Scene Modeling . . . . . . . . . . . . . . . . . . . . . 396
12.3.2 Path Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
12.3.3 Automatic Activity Analysis . . . . . . . . . . . . . . . . . . . . 403
12.4 Spatial-Temporal Action Classification and Recognition . . . . . . 405
12.4.1 Motion Classification . . . . . . . . . . . . . . . . . . . . . . . . . 405
12.4.2 Action Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . 407
12.5 Key Points and References for Each Section . . . . . . . . . . . . . . . 414
Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418

Appendix A: Visual Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419

A.1 Shape Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
A.2 Spatial Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
A.2.1 Nonvisual Indices of Depth . . . . . . . . . . . . . . . . . . . . . 421
A.2.2 Binocular Indices of Depth . . . . . . . . . . . . . . . . . . . . . 422
A.2.3 Monocular Indices of Depth . . . . . . . . . . . . . . . . . . . . . 424
Contents xiii

A.3 Motion Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427

A.3.1 The Condition of Motion Perception . . . . . . . . . . . . . . . 427
A.3.2 Detection of Moving Objects . . . . . . . . . . . . . . . . . . . . 429
A.3.3 Depth Motion Detection . . . . . . . . . . . . . . . . . . . . . . . 429
A.3.4 Real Motion and Apparent Motion . . . . . . . . . . . . . . . . 430
A.3.5 Correspondence Matching of Apparent Motion . . . . . . . 430
A.3.6 Aperture Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
A.3.7 Dynamic Indices of Depth . . . . . . . . . . . . . . . . . . . . . . 433
A.4 Key Points and References for Each Section . . . . . . . . . . . . . . . 434
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
Answers to Self-Test Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
Chapter 1 Computer Vision Overview . . . . . . . . . . . . . . . . . . . . . . . . 435
Chapter 2 Camera Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
Chapter 3 Three-Dimensional Image Acquisition . . . . . . . . . . . . . . . . . 436
Chapter 4 Video Image and Motion Information . . . . . . . . . . . . . . . . . 437
Chapter 5 Moving Object Detection and Tracking . . . . . . . . . . . . . . . . 437
Chapter 6 Binocular Stereo Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . 438
Chapter 7 Monocular Multiple Image Recovery . . . . . . . . . . . . . . . . . 439
Chapter 8 Monocular Single Image Recovery . . . . . . . . . . . . . . . . . . . 439
Chapter 9 Three-Dimensional Scene Representation . . . . . . . . . . . . . . 440
Chapter 10 Scene Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
Chapter 11 Knowledge and Scene Interpretation . . . . . . . . . . . . . . . . . 441
Chapter 12 Spatial-Temporal Behavior Understanding . . . . . . . . . . . . . 442
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
Chapter 1
Computer Vision Overview

Computer vision is an information subject/discipline that uses computers to realize

the functions of human vision system (HVS). This book mainly introduces the high-
level content of computer vision, which can be used as a textbook for in-depth study
of computer vision. Readers can read this book after studying 2D Computer Vision:
Principles, Algorithms and Applications.
The process of human vision (in brief, vision) can be regarded as a complex
process from sensation (obtaining the image by the 2-D projection of the 3-D world)
to perception (gaining the content and meaning of the 3-D world from the 2-D
image) process. The ultimate goal of vision in a narrow sense is to make a mean-
ingful explanation and description of the scene for the observer, and in a broad sense,
it is also to make a behavior plan based on these explanations and descriptions, as
well as the surrounding environment and the wishes of the observer. Computer
vision is artificial vision or human-made vision. It is to use computers to realize
human visual functions, hoping to make meaningful judgments about actual targets
and scenes based on the perceived images.
The sections of this chapter are arranged as follows:
Section 1.1 introduces the characteristics of human vision, the brightness properties
of vision, the spatial properties of vision, and the temporal properties of vision, as
well as makes some discussions on visual perception.
Section 1.2 discusses the research purpose, research tasks, and research methods of
computer vision. It also introduces the visual computational theory proposed by
Marr in more detail. Moreover, a combined presentation for some improvement
ideas is provided.
Section 1.3 gives a general introduction to the 3-D vision system that obtains 3-D
spatial information and realizes the understanding of the scene. It compares and
discusses the layers of computer vision and image technology, leading to the
main content of this book.
Section 1.4 presents the structure of the book and gives the brief summaries of each
chapter.

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 1
Y.-J. Zhang, 3-D Computer Vision, https://doi.org/10.1007/978-981-19-7580-6_1
2 1 Computer Vision Overview

1.1 Human Vision and Characteristics

Computer vision (also called artificial vision or human-made vision) is developed on

the basis of human vision. There are two meanings here. One is that computer vision
needs to realize the functions of human vision, so it could imitate the system
structure and functional modules of human vision; the other is that computer vision
needs to extend the functions of human vision, so it could use the features of human
vision to improve the efficiency and effectiveness of these functions.
The following first gives an overview of the functional characteristics of human
vision and then discusses some important visual brightness, spatial, and temporal
properties.

1.1.1 Visual Characteristics

First compare vision with some related concepts.

1.1.1.1 Vision and Other Sensations

It is generally believed that humans have five senses of vision, hearing, smell, taste,
and touch and the corresponding sensory organs, for obtaining information from the
objective world. Among them, vision provides humans with most of the data they
receive; in other words, humans often rely more on vision than other senses when
they acquaint the world. For example, the input information obtained by humans
from eye observation often reaches several million bits, and the data rate during
continuous viewing can exceed tens of million bits per second. The human brain has
more than 1010 cells/neurons, some of which have more than 10,000 connections
(or synapses) with other neurons. It is estimated that the amount of visual informa-
tion the brain receives from the eyes is at least two orders of magnitude larger than all
the information obtained from other sense organs.

1.1.1.2 Vision and Computer Vision

Although computer vision needs to realize the functions of human vision, there are
still differences between human vision and computer vision. Human vision first
receives light stimulation in a certain wavelength range from the external environ-
ment through the sensory organs (eyes) of the visual system to the retina and then
encodes and processes the perceptive organs (brain or cerebral visual cortex) of the
visual system to obtain subjective feelings. Therefore, vision involves not only
physics and chemistry but also psychophysiology. Computer vision mainly relies
on photoelectric conversion for image acquisition, then uses processing and analysis
1.1 Human Vision and Characteristics 3

functions to obtain objective data, and makes stricter reasoning and judgments based
on these results.

1.1.1.3 Vision and Machine Vision

In the early days, computer vision put more emphasis on the research of vision
science and design system software, while machine vision not only considered
design system and software but also considered hardware environment and image
acquisition technology as well as the integration of vision systems. Therefore, from
the perspective of vision system integration, machine vision systems are more
comparable to human vision systems (human visual systems). However, with the
development of electronic technology and computer technology, true and real-time
applications can be realized to a considerable extent on a stand-alone PC. Due to the
increase in knowledge in related fields, the difference between machine vision and
computer vision has been significantly reduced, and they are more interchangeably
used nowadays.

1.1.1.4 Vision and Image Generation

Vision can be regarded as a process of obtaining description and explanation of the

scene from the image of the scene with the help of the knowledge of the law of image
formation. The image generation in graphics can be seen as the process of generating
images from the abstract description of the scene with the help of the knowledge of
the law of image formation. Although they have some parallels/similarities, certain
people regard them as inverse processes, but their complexity is quite different. The
image generation process is completely deterministic and predictable, and the visual
process involves not only providing a list of all possible explanations but also
providing the most likely interpretation. This search process is one to many and
may be accompanied by combinatorial explosion. Therefore, vision is inherently
more complex than image generation in graphics.

1.1.2 Brightness Properties of Vision

The brightness of vision corresponds to the light intensity that the human eye feels
from the scene. A psychological term closely related to brightness is subjective
brightness or subjective luminance The subjective brightness refers to the bright-
ness of the observed object, which is judged by the human eyes according to the
intensity of the light stimulation of the retina. Examples of three typical brightness
properties in which the perception of brightness is related to multiple factors are as
follows.
4 1 Computer Vision Overview

1.1.2.1 Simultaneous Contrast

The subjective brightness felt from the surface of an object is not only related to the
brightness of the surface itself but also to the relative relationship (ratio) between the
brightness of the surface and the surrounding environment (background). If two
objects with different brightness have a similar relative relationship with their
respective backgrounds, they can appear to have the same brightness. At this time,
the subjective brightness perceived by people has nothing to do with the absolute
value of the object brightness. Conversely, the surface of the same object will appear
brighter if it is placed on a darker background and will appear darker if it is placed on
a brighter background. This phenomenon is called simultaneous contrast, also
called conditional contrast.
Example 1.1 Simultaneous Contrast Example
All the small squares in the center of big squares (background) in Fig. 1.1 have
exactly the same brightness. However, it looks brighter when it is on a dark
background and looks darker when it is on a bright background. So, it feels like
when these four pictures are viewed from left to right, the small square in the center
gradually darkens. This is just the result of simultaneous contrast.

1.1.2.2 Mach Band Effect

At the boundary of the touch regions with different brightness in the object, human
vision may estimate the brightness value too high or too low. In other words, the
subjective brightness felt from the surface of an object is not a simple proportional
function of the illuminance the object receives. This phenomenon was discovered by
Mach, so it is called the Mach band effect.
Example 1.2 Example of Mach Band Effect
Figure 1.2a is a Mach band pattern, which includes three parts: the left side is a
uniform low-brightness area, the right side is a uniform high-brightness area, and the
middle is a gradual transition from low brightness to high brightness. Fig. 1.2b
shows the actual brightness distribution (three straight lines) from left to right. If one
looks at Fig. 1.2a with the eyes, one will find that there is a darker band at the
junction of the left and middle bands than the left band and a darker band at the
junction of the middle and right bands than the right band. The subjective brightness

Fig. 1.1 Example of simultaneous contrast

1.1 Human Vision and Characteristics 5

Fig. 1.2 Schematic

diagram of Mach band effect
(a)

Brightness

(b)
Distance

Subjective
brightness
(c)
Distance

obtained should be as shown in Fig. 1.2c. In fact, the darker band and the brighter
band do not exist objectively; it is the result of subjective brightness perception.

1.1.2.3 Contrast Sensitivity

Contrast sensitivity (also called contrast susceptibility) reflects the ability of the
human eye to distinguish differences in brightness. It is affected by the observation
time and the size of the observed object. If a grid composed of lines of different
thickness and contrast is used for testing, the closer the contrast between the bright
and dark lines of the grid perceived by the eyes is to the contrast between the bright
and dark lines of the original test grid, the greater the contrast sensitivity. Under ideal
conditions, people with good eyesight can distinguish a brightness contrast of 0.01,
which means that the contrast sensitivity can reach up to 100.
If the abscissa represents the thickness of the bright and dark lines of the test grid
and the ordinate represents the contrast sensitivity, then the measured result will give
the modulation transfer function of the visual system, that is, the ability that human
visual system accurately converts the test grid into an optical image. Here the grid
thickness can be expressed by spatial frequency, and its unit is the number of circles
(number of lines) contained in each degree of viewing angle, that is, circle/degree
(CPD).
The contrast sensitivity can be regulated by the light modulation factor M. If Lmax,
Lmin, and Lav are, respectively, the maximum, minimum, and average brightness
values, then
 Visit https://ebookmass.com
now to explore a rich
collection of eBooks and enjoy
exciting offers!
6 1 Computer Vision Overview

L max - L min
M= ð1:1Þ
Lav

1.1.3 Spatial Properties of Vision

Vision is first and foremost a spatial experience, so the spatial properties of vision
have a great influence on visual effects.

1.1.3.1 Spatial Cumulative Effect

Vision has a cumulative effect in space. The range of light stimulus intensity
experienced by the human eye can reach about 13 orders of magnitude. If described
by light illuminance, the lowest absolute stimulus threshold is 10-6 lx (lux), and the
highest absolute stimulus threshold exceeds 107lx. Under the best conditions, every
light quantum in the peripheral region of the retina will be absorbed by a rod cell, and
a visual response can be caused by only a few light quanta. This is considered to have
taken place in a complete spatial accumulation, and it can be described by the law of
the inverse ratio of light intensity and area. This law can be written as

E c = kAL ð1:2Þ

where Ec is the absolute threshold of vision, which is the critical light energy
required for 50% detection probability (i.e., the light energy when the light stimu-
lation is observed once in every two tests in multiple experiments); A is the
cumulative area; L is the brightness; and k is a constant, which is related to the
units used by Ec, A, and L. Note that the area that can satisfy the above law has a
critical value Ac (corresponding to a round solid angle with a diameter of about
0.3 rad). When A < Ac, the above law holds; otherwise the above law does not hold.
It can be seen that the spatial cumulative effect can be understood as follows:
when a small and weak light spot is presented alone, it may be invisible (cannot
cause a visual response), but when multiple such light spots are connected together
as a large spot at the same time, one can see it. Its functional significance lies in: large
objects may be seen in a dark environment even if the outline is blurred.

1.1.3.2 Spatial Frequency

The spatial frequency corresponds to the changing speed of the visual image in
space. This can be tested with stripes whose brightness sinusoidally changes in
space. The brightness function Y(x, y) = B(1 + mcos2πfx). Among them, B is the
basic brightness, m is the amplitude (corresponding to black and white contrast), and
1.1 Human Vision and Characteristics 7

f is the fringe frequency (corresponding to the fringe width). The spatial resolution
capability can be tested by changing m when f is given as a fixed value. Obviously,
the larger m, the stronger the spatial resolution ability. In practice, to test the
minimum m value that can distinguish bright and dark stripes at different angles
and frequencies, one can define 1/m minute (1′) as the contrast sensitivity. Usually
the human eye, in terms of spatial frequency felling, is equivalent to a band-pass
filter (more sensitive to the intermediate thickness of the stripes), the most sensitive
is 2 ~ 5 CPD, and the spatial cut-off frequency is 30 CPD.
When a person observes a still image, the eyeball is not still in one place, usually
after staying in one place for a few hundred milliseconds to complete the image
acquisition, the eyeball will move to another place to take another image, and so
on. This kind of movement is called saccadic eye movement. Studies have shown
that jumping movement can increase contrast sensitivity, but the peak sensitivity will
decrease.

1.1.3.3 Visual Acuity

Visual acuity is usually defined as the reciprocal of the viewing angle value
corresponding to the smallest detail that can be distinguished under certain condi-
tions. The smaller the viewing angle, the greater the visual acuity. If V is used for
visual acuity, then V = 1/(viewing angle). It represents the ability of the human eye
to correctly distinguish the details and contours of objects. A visual acuity of
1 indicates the resolution capability at a standard distance when the corresponding
viewing angle is 1°. The actual viewing angle of the human eye is 30″ ~ 60″ (this is
basically consistent with the cone cell diameter of about 0.004 mm), that is, the best
visual acuity can reach 2.0.
Visual acuity is affected by many factors, including the following:
1. Distance: When the distance of the object from the observer increases, the visual
acuity of the human eye decreases. This phenomenon is most obvious at about
10 m, and beyond a certain distance limit, the details of the object can no longer
be recognized.
2. Brightness: Increasing the brightness of the object (or enlarging the pupil) will
increase the visual acuity. The relationship between visual acuity and brightness
I is

V = a log I þ b ð1:3Þ

where a and b are constants. Visual acuity increases with increasing brightness, and
the relationship between the two is logarithmic. If the brightness continues to
increase to a certain level, the visual acuity will be close to saturation and will not
increase.
Random documents with unrelated
content Scribd suggests to you:
 BLACK-THROATED DIVER.

Colymbus Arcticus, Linn.

PLATE CCCXLVI. Male, Female, and Young.

One of the most remarkable circumstances relative to this beautiful

bird, which is intermediate between the Red-throated Diver and the
Loon, is the extraordinary extent to which the wanderings of the
young are carried in autumn and winter. It breeds in the remote
regions of the north, from which many of the old birds, it would seem,
do not remove far, while the young, as soon as they are able to
travel, take to wing and disperse, spreading not only over the greater
part of the United States, but beyond their south-western limits. In
the Texas I saw individuals of this species as late as the middle of
April 1837; and I find it enumerated in a list of the birds observed by
my young friend Dr J. K. Townsend on the Columbia River, where
he also met with Columbus glacialis. Its ramblings over a
considerable portion of northern and eastern Europe have equally
been noted, and it has been found breeding in the extreme north of
Scotland.
For many years I knew the young of this bird only by the name
“Imber Diver,” applied by Bewick to that of another species, and now
have pleasure in looking upon a drawing of mine, made about thirty
years ago, with that appellation attached to it. Very few old birds in
full plumage have been procured within the limits of the United
States, and none in as far as I know, farther south than the Capes of
Delaware.
No sooner has the foliage of the trees that border our western waters
begun to drop and float on the gentle current of the fair Ohio, than
the Black-throated Diver makes its appearance there, moving slowly
with the stream. The Mississippi, Missouri, and their tributaries, are
at the same period supplied with these birds. Along our eastern and
southern shores they are seen from the end of autumn until spring.
Whilst in Labrador, I saw a few pairs courting on wing, much in the
manner of the Red-throated Diver; but all our exertions failed to
procure any of the nests, which I therefore think must have been
placed farther inland than those of the Loon or Red-throated Diver. I
observed however, that in their general habits they greatly resemble
those species, for on alighting on the water, they at once immerse
their bills, as if for the purpose of ascertaining whether it yields a
supply of suitable food, and afterwards raise themselves and beat
their wings.
This species has almost as powerful a flight as the Great Northern
Diver or Loon, and I think shoots through the air with even greater
velocity. When flying it moves its wings rapidly and continuously, and
has the neck and feet stretched out to their full length. I well recollect
that while I was standing near the shore of a large inlet in South
Carolina, one of these birds, being shot while passing over my head
at full speed, did not, on account of the impetus, reach the ground
until upwards of twenty yards beyond me. They are equally expert at
diving, and fully as much so in eluding the pursuit of their enemies
when wounded. I saw my friend Mr Harris bring down one from on
wing, on which Napoleon Coste, and William Taylor, Captains of
the Revenue Cutter and Tender of which we had the use, paddled in
pursuit of it in a light canoe; but, although they advanced with all the
address of Indians, they proved unsuccessful, for after following it
both in the Bay of Cayo Island, and in the Bay of Mexico, for nearly
an hour, they were obliged to return without it, having found it
apparently not in the least fatigued, although it had dived sufficiently
often to travel above two miles, shifting its course at each immersion.
It is curious to observe how carefully these birds avoid the danger of
sudden storms or heavy gales. On such occasions, I have seen
Divers at once seek the lee of rocks, islands, or artificial
embankments, where they could not only remain in security, but also
procure their accustomed food. At other times, when striving against
the tempest, they dive headlong from on wing, and are sure to
reappear in the smooth parts which sailors term the trough.
I once caught one of these birds on the Ohio, it having been
incapacitated from diving by having swallowed a large mussel, which
stuck in its throat. It was kept for several days, but refused food of
every kind, exhibited much bad humour, struck with its bill, and died
of inanition. The food of this species consists of fish, aquatic reptiles,
testaceous mollusca, and all sorts of small crustaceous animals. Its
flesh resembles that of the Loon, and is equally unfit to be eaten.
The eggs, which are sometimes two, more frequently three, average
three inches in length, by two in their greatest breadth, which is
about a third of the whole length distant from the extremity. Their
form is that of the Red-throated Diver, which however they exceed in
size. The shell is rather thick, the surface roughish, the ground
colour chocolate tinged with olive, sparingly spotted at the larger end
with very dark umber and black, and sprinkled all over with very
small dots of the same colour.
I have represented an adult male, a female, and a young bird.

Colymbus arcticus, Linn. Syst. Nat. vol. i. p. 221.—Lath. Ind. Ornith. vol. ii.
p. 800.—Ch. Bonaparte, Synopsis of Birds of United States, p. 420.
Colymbus arcticus, Black-throated Diver, Richards. and Swains. Fauna
Boreali-Americana, vol. ii. p. 475.
 Black-throated Diver, Nuttall, Manual, vol. ii.

Adult Male. Plate CCCXLVI. Fig. 1.

Bill as long as the head, straight, stout, higher than broad at the
base, much compressed toward the end, and tapering to a point.
Upper mandible with the dorsal line descending and considerably
convex toward the end, the ridge convex, narrowed toward the point,
the sides convex beyond the nostrils, the edges involute for half their
length in the middle, direct at the base and toward the end, the tip
narrow and sharpish. Nasal groove rather long and narrowed;
nostrils sub-basal, linear, direct, pervious. Lower mandible with the
angle extremely narrow, and very long, the dorsal line ascending and
very slightly convex, the ridge convex and narrow, the edges sharp
and involute, the tip attenuated.
Head of moderate size, oblong, narrowed before. Neck rather long
and thick. Eyes of moderate size. Body elongated, much depressed,
of an elliptical form viewed from above. Wings small. Feet short,
rather large, placed very far back; tibia almost entirely concealed;
tarsus short, exceedingly compressed, sharp-edged before and
behind, covered all over with reticulated angular scales, hind toe
extremely small, externally marginate, connected with the second for
half its length by a membrane, which extends, narrowing, to the end;
the anterior toes connected by articulated membranes, the fourth or
outer longest, the third a little shorter, the second considerably
shorter than the third; all covered above with numerous narrow
scutella; the second toe with a free two-lobed membrane, the claws
very small, depressed, blunt.
Plumage short and dense, of the head and neck very short, soft and
blended; of the lower parts short, blended, stiffish, considerably
glossed; of the upper compact, glossy; the feathers on the lower part
of the sides of the neck much incurved, oblong with the terminal
barbs stiff; those of the fore part of the back and the scapulars
straight, oblong, abrupt. Wings proportionally very small and narrow,
curved; primaries strong, tapering, the first longest, the second
slightly shorter, the rest rapidly graduated; secondaries very short,
broad, and rounded. Tail extremely short, rounded, of eighteen
feathers.
Bill black. Iris deep bright red. Feet greyish-blue, their inner sides
tinged with yellow; claws black, that of the inner toe yellowish at the
base. The upper part of the head and the hind neck are light grey or
hoary, the fore part and sides of the head darker. The upper parts
are glossy black tinged with green anteriorly, and shaded with brown
behind. On the fore part of the back are two longitudinal bands of
transverse white bars, the feathers being tipped with that colour; the
scapulars, excepting the outer, are marked in the same manner with
transverse rows of rather large square spots. Most of the wing-
coverts have two roundish spots of white near the end. The quills are
blackish-brown, tinged with grey externally, paler on the inner webs;
the tail also blackish-brown. The fore neck, to the length of six and a
half inches, is purplish-black, ending angularly below, and with a
transverse interrupted band of linear white spots near the upper part;
beyond which the sides of the neck are blackish-brown, with several
longitudinal white streaks, formed by the edges of the feathers; on
the lower part of the neck a broad space is occupied by these
longitudinal, dusky, and white streaks the former of which gradually
become narrower. The lower parts are pure white, excepting a
longitudinal band on the sides under the wing, which is dusky.
Length to end of tail 29 inches, to end of wings 27 1/2, to end of
claws 33; extent of wings 39 1/2; wing from flexure 12 3/4; tail 2 3/4;
bill along the ridge 2 5 1/2/12, along the edge of lower mandible
3 4 1/2/12; tarsus 3 1/12; hind toe 8/12, its claw 2/12; second toe 3 2/12,
 1/2 1/2
its claw 5 /12; third toe 3 8/12, its claw 5 5 /12; fourth toe 4 1/4, its
claw 4 1/2/12.

Adult Female. Plate CCCXLVI. Fig. 2.

The Female is smaller than the male, but is similarly coloured.
Young in Winter. Plate CCCXLVI. Fig. 3.
The texture of the plumage is less dense, the feathers on the neck
being more downy, and those of the back oblong and rounded. The
bill is light bluish-grey, dusky along the ridge; the iris brown; the feet
more dusky. The upper part of the head and the hind neck are dark
greyish-brown; the sides of the head greyish-white, minutely
streaked with brown. The upper parts have a reticulated or scaly
appearance, the feathers being brownish-black, with broad bluish-
grey margins; the rump dull brownish-grey. The primaries and their
coverts are brownish-black, the secondaries and tail-feathers dusky,
margined with grey. The fore part of the neck is greyish-white,
minutely and faintly dotted with brown, its sides below streaked with
the same; the lower parts, including the under surface of the wing,
pure white; the sides of the body and rump, with part of the lower
tail-coverts, dusky, edged with bluish-grey.
When in their first downy plumage, the young are of a uniform
brownish-black colour.
 SMEW OR WHITE NUN.

Mergus Albellus, Linn.

PLATE CCCXLVIL. Male and Female.

The Smew is a bird of extremely rare occurrence in the United

States, insomuch that it must be considered merely as a transient or
accidental visitor. Indeed I have felt strong misgivings on reading
Wilson’s article on this species, and cannot but think that he is
mistaken when he states that it “is much more common on the coast
of New England than farther south,” and again “In the ponds of New
England, and some of the lakes in the State of New York, where the
Smew is frequently observed—.” Now, although I have made diligent
inquiry, not only in New England, but in every part of our country
where I thought it likely that the Smew might occur, I have not met
with any person well acquainted with birds of this family, who has
seen it. Wilson, in short, was in all probability misinformed, and it is
my opinion that his figure was made from a stuffed European
specimen which was then in Peale’s Museum in Philadelphia, and
that he had taken the Buffel-headed Duck, seen at a distance, for
this species, as I am aware has been the case with other individuals.
The only specimen procured by me was shot by myself on Lake
Barataria, not far from New Orleans, in the winter of 1819. It was an
adult female in fine plumage. How it had wandered so far south is an
enigma to me; but having found it, and made a drawing of it on the
spot, I have taken the liberty to add one of the other sex from an
equally fine specimen. After all, the Smew can scarcely be
considered as belonging to the American Fauna, any more than our
Fork-tailed Hawk can with propriety be called a denizen of England;
and in this I am supported by all the great navigators of our Arctic
Seas, such as Ross, Parry, and Franklin, none of whom, nor any
of their companions, ever met with a single individual of this beautiful
bird.

Mergus Albellus, Linn. Syst. Nat. vol. i. p. 209.—Lath. Ind. Ornith. vol. ii. p.
831.—Ch. Bonaparte, Synopsis of Birds of United States, p. 398.
Smew or White Nun, Mergus Albellus, Wils. Amer. Ornith. vol. viii. p. 126,
pl. 71, fig. 4. Male.
The Smew, or White Nun, Nuttall, Manual, vol. ii. p. 467.

Adult Male Plate CCCXLVII. Fig. 1.

Bill rather shorter than the head, straight, rather slender, a little
higher than broad at the base, tapering, somewhat cylindrical toward
the end. Upper mandible with the dorsal outline sloping gently and
slightly concave to the middle, then straight, at the tip declined, the
ridge rather broad and flat at the base, then convex, the sides
sloping at the base, convex toward the end, the edges serrate
beneath, with about forty slightly reversed, compressed, tapering,
tooth-like lamellæ, the unguis elliptical, much curved. Nasal groove
oblong, sub-basal, filled by a soft membrane; nostrils oblong,
submedial, direct, pervious. Lower mandible with the angle very
narrow and extended to the obovate, very convex unguis, the sides
rounded, with a long groove, the edges with about sixty
perpendicular sharp lamellæ.
Head of moderate size, oblong, compressed. Neck of moderate
length. Body full and depressed. Feet placed far behind, extremely
short; tibia bare for a quarter of an inch; tarsus extremely short,
much compressed, anteriorly covered with a series of very small
scutella, and another row on the lower half externally, the sides
reticulate. Hind toe very small, with an inferior free membrane;
anterior toes double the length of the tarsus; the second shorter than
the fourth, which is nearly as long as the third; all connected by
reticulated webs, of which the outer is deeply emarginate. Claws
short, considerably curved, compressed, acute, that of the middle
toe with a thin inner edge.
Plumage full, soft, and blended; feathers of the head and upper part
of the hind neck very slender, and elongated along the median line
into a narrow decurved crest; those of the shoulders obovate and
abrupt, of the rest of the upper parts ovate, of the lower elliptical.
Wings very short, narrow, curved, and pointed; primaries narrow,
tapering, the first scarcely longer than the second, the rest rapidly
graduated; secondaries short, narrow, rounded, the inner tapering to
an obtuse point. Tail short, graduated, of sixteen rather narrow,
tapering feathers.
Bill dark greyish-blue. Iris bright red. Feet livid blue, claws dusky.
The general colour of the plumage is pure white; a short band on
each side of the hind neck bordering the crest, duck-green; a broad
patch on the lore and below the eye, a narrow band across the lower
part of the hind neck, formed by single bars near the tips of the
feathers, the middle of the back in its whole length, a short
transverse bar under the fore edge of the wing, the anterior margin
of that organ to beyond the carpal joint, the outer edges of the
scapulars, the primary coverts, the secondary coverts, and the outer
secondary quills, excepting the tips of both, deep black. The quills
are also black, but of a less deep tint; the hind part of the back
becomes tinged with grey, and the rump and tail-feathers are dusky
grey. The sides of the body and rump are white, finely undulated with
blackish-grey.
Length to end of tail 17 1/2 inches, to end of claws 18 1/4, to end of
1/2
wings 15 1/2; extent of wings 27; bill along the ridge 1 3 /12, along
1/
the edge of lower mandible 1 7 2/12; wing from flexure 7 3/4; tail 3 1/2;
1/
tarsus 1 1 /12; first toe 1/2, its claw 2/12; second toe 1 1/2, its claw
2

4/ ,
12third toe 1 11/12, its claw 4 1/2/12; fourth toe 1 10/12, its claw 5/12.
Weight 1 lb. 8 oz.
Adult Female. Plate CCCXLVII. Fig. 2.
The Female is much smaller. The feathers of the hind part of the
head and neck are also elongated so as to form a crest. The bill, iris,
and feet, are coloured as in the male. All the lower parts are white,
excepting a broad band of light grey across the middle of the neck,
and a narrow portion of the sides, which are of a deeper tint. There is
a patch of brownish-black on the lore and beneath the eye; the upper
part of the head and half of the hind neck, are light reddish-brown;
the rest of the hind neck, and all the upper parts, bluish-grey, darker
behind, and in the middle of the back approaching to black. The
wings as in the male, that is black, with a large patch of white, and
two narrow transverse bands of the same; the tail dusky grey.
Length to end of tail 15 1/4 inches, to end of claws 16 1/2, to end of
wings 14 1/2; extent of wings 25. Weight 1 lb. 4 oz.
 GADWALL DUCK.

Anas strepera, Linn.

PLATE CCCXLVIII. Male and Female.

I have met with this species along the whole of our Atlantic coast,
from Eastport in Maine to Texas. It is, however, more abundant in the
interior than in most of our maritime districts, and is particularly so on
the tributaries of the Ohio, Missouri, and Mississippi. In the early part
of autumn and late in spring many are found on the margins of our
great lakes. Yet the Gadwall has been represented as not plentiful in
the United States, probably on account of its being generally
dispersed, and not congregated in particular districts.
The Creoles of Louisiana name it “Violon,” on account of the
whistling sound of its wings. It arrives in the neighbourhood of New
Orleans and the mouths of the Mississippi along with the Widgeon,
and is fond of the company of the Red-head, to which it is about
equal as an article of food. The Gadwalls are usually seen in small
flocks, and during winter resort to the larger lakes and the pools in
the interior of the great marshes, adjoining the waters of the Gulf. In
that part of the country they feed on small fish, insects, and aquatic
grasses. Fewer of them are found in Massachusetts and the State of
New York than elsewhere, and this probably on account of these
districts being more elevated and less marshy than those farther
south. My friend Dr Bachman informs me that they are rather
plentiful in South Carolina, where they are considered good eating,
and where they arrive in the beginning of October, but are more
frequently met with at that season, and in early spring, than during
winter, when a single individual may sometimes be seen in a flock of
other ducks.
While we were in the Texas, in the latter part of April and the
beginning of May, we found the Gadwall quite abundant on all the
inland ponds and streams, as well as on the brackish pools and
inlets of the islands and shores of Galveston Bay. Many of them had
paired and separated from the other ducks; and I was assured that
this species breeds there, as does the Dusky Duck, the Mallard, the
Blue-winged Teal, the Widgeon, and the Shoveller, the young of all
these species being plentiful in the end of June and beginning of
July. I was satisfied as to the truth of the repeated assurances I had
received on this subject, by observing the manners of individuals of
all these species before my departure from that country. After a
continuance of rainy weather, Gadwalls are found in great numbers
on the vast prairies of Oppelousas and Attacapas, where I have
been told they continue until very late in spring, and some remain to
breed.
This species dives well on occasion, especially on being wounded.
At the appearance of danger, it rises on wing—whether from the
ground or from the water—at a single spring, in the manner of the
Mallard, and, like it also, ascends almost perpendicularly for several
yards, after which it moves off in a direct course with great celerity. I
have never seen it dive on seeing the flash of the gun; but when
approached it always swims to the opposite part of the pond, and,
when the danger increases, flies off. On being wounded, it
sometimes by diving makes its escape among the grass, where it
squats and remains concealed. It walks with ease, and prettily, often
making incursions upon the land, when the ponds are not
surrounded by trees, for the purpose of searching for food. It nibbles
the tender shoots and blades of grasses with apparent pleasure, and
will feed on beech-nuts, acorns, and seeds of all kinds of gramineæ,
as well as on tadpoles, small fishes, and leeches. After rain it alights
in the corn-fields, like the Mallard, and picks up the scattered grains
of maize. The common notes or cry of the female have a
considerable resemblance to those of the female Mallard; but the cry
of the male is weaker as in that species.
It is by no means shy in the Western Country, where I have often
found it associating with other species, which would leave the pond
before it. Near the sea, however, it is much more wary, and this no
doubt on account of the greater number of persons who there follow
shooting as a regular and profitable employment. From the following
note of my friend Dr Bachman, you may judge how easily this fine
species might be domesticated.
“In the year 1812 I saw in Dutchess County, in the State of New
York, at the house of a miller, a fine flock of ducks, to the number of
at least thirty, which, from their peculiar appearance, struck me as
differing from any I had before seen among the different varieties of
the tame Duck. On inquiry, I was informed that three years before, a
pair of these ducks had been captured in the mill pond, whether in a
trap, or by being wounded, I cannot recollect. They were kept in the
poultry-yard, and, it was said, were easily tamed. One joint of the
wing was taken off, to prevent their flying away. In the following
spring they were suffered to go into the pond, and they returned daily
to the house to be fed. They built their nest on the edge of the pond,
and reared a large brood. The young were perfectly reconciled to
domestication, and made no attempts, even at the migratory season,
to fly away, although their wings were perfect. In the following
season they produced large broods. The family of the miller used
them occasionally as food, and considered them equal in flavour to
the common duck, and more easily raised. The old males were more
beautiful than any that I have examined since; and as yet
domestication had produced no variety in their plumage.”
The migration of this species extends to the Fur Countries, where it
is said to breed. The description of a male killed on the
Saskatchewan River, on the 22d of May 1827, is given in the Fauna
Boreali-Americana; and I have a fine male procured by Dr
Townsend on the Columbia River.

Anas strepera, Linn. Syst. Nat. vol. i. p. 200.—Lath. Ind. Ornith. vol. ii. p.
859.—Ch. Bonaparte, Synopsis of Birds of United States, p. 383.
Gadwall, Anas strepera, Wils. Amer. Ornith. vol. viii. p. 120, pl. 71, fig. 1.—
Richards. and Swains. Fauna Bor.-Amer. vol. ii. p. 440.
Gadwall or Grey, Nuttall, Manual, vol. ii. p. 383.

Adult Male. Plate CCCXLVIII. Fig. 1.

Bill nearly as long as the head, deeper than broad at the base,
depressed towards the end, the sides parallel, the tip rounded.
Upper mandible with the frontal angles short and obtuse, the dorsal
line at first sloping, then slightly concave and direct, the ridge broad
and flat at the base, then broadly convex, the edges soft, with about
fifty internal lamellæ, the unguis roundish, curved abruptly at the
end. Nostrils sub-basal, lateral, rather small, oblong, pervious. Lower
mandible flattened, its angle very long and narrow, the dorsal line
very short, slightly convex, the edges soft, with about sixty lamellæ.
Head of moderate size, oblong, compressed. Neck rather long,
slender. Body elongated, slightly depressed. Feet very short; tibia
bare for about a quarter of an inch; tarsus very short, compressed,
anteriorly with two series of scutella, the outer shorter, the rest
covered with reticulated angular scales; toes obliquely scutellate
above; first very small, free, with a narrow membrane beneath; third
longest, fourth considerably shorter, second shorter than fourth, their
connecting webs entire, on the edge crenate; the second or inner toe
with a membranous margin. Claws small, slightly arched,
compressed, rather acute, the hind one very small and more curved,
that of the middle toe with an inner sharp edge.
Plumage dense, soft, blended. Feathers of the head short, of the
occiput and nape a little elongated, of the lower parts glossy with the
extremities of the filaments stiffish. Wings rather long, little curved,
pointed; the first quill longest, the rest rapidly graduated; secondaries
very broad, but pointed, the inner much elongated, and tapering to a
point. The tips of the filaments of the outer web of the first primary
are separated and curved a little forwards. Tail short, rounded, of
sixteen strong pointed feathers, of which the middle pair project
considerably.
Bill bluish-black. Iris reddish hazel. Feet dull orange-yellow, claws
brownish-black, webs dusky. Head light yellowish-red, the upper part
and nape much darker and barred with dusky; the rest dotted with
the same. The lower part of the neck, the sides of the body, the fore
part of the back, and the outer scapulars, undulated with dusky, and
yellowish-white, the bands much larger and semicircular on the fore
part of the neck and breast; the latter white, the abdomen faintly and
minutely undulated with brownish-grey; the elongated scapulars
brownish-grey, broadly margined with brownish-red; the hind part of
the back brownish-black; the rump all round, and the upper and
lower tail-coverts, bluish-black. The anterior smaller wing-coverts are
light grey, undulated with dusky, the middle coverts of a deep rich
chestnut-red; primary coverts brownish-grey, outer secondary
coverts darker and tinged with chestnut, the rest black, excepting the
inner, which are grey. Primaries and inner elongated secondaries
brownish-grey, of which colour also are the inner webs of the rest,
part of the outer webs of five of the outer black, and their terminal
margins white, of which colour are the whole outer webs of the three
next to the inner elongated quills. Tail brownish-grey, the feathers
margined with paler.
Length to end of tail 21 3/4 inches, to end of wings 19, to end of
claws 23 1/4; extent of wings 35; bill along the ridge 1 3/4, along the
edge of lower mandible 1 7/8; wing from flexure 11; tail 4 3/8; tarsus
1 1/2; hind toe and claw 1/2; second toe 1 5/8, its claw 4/12; third toe
1 7/8, its claw 4/12; outer toe 1 7 1/2/12, its claw 2/12. Weight 1 lb. 10 oz.

Adult Female. Plate CCCXLVIII. Fig. 2.

The female is considerably smaller. Bill dusky along the ridge, dull
yellowish-orange on the sides. Iris hazel. Feet of a fainter tint than in
the male. Upper part of head brownish-black, the feathers edged
with light reddish-brown; a streak over the eye, the cheeks, the
upper part of the neck all round, light yellowish-red tinged with grey,
and marked with small longitudinal dusky streaks, which are fainter
on the throat, that part being greyish-white; the rest of the neck, the
sides, all the upper parts and the lower rump feathers brownish-
black broadly margined with yellowish-red. Wing-coverts brownish-
grey, edged with paler; the wing otherwise as in the male, but the
speculum fainter. Tail-feathers, and their coverts dusky, laterally
obliquely indented with pale brownish-red, and margined with
reddish-white.
Length to end of tail 19 1/4 inches, to end of wings 18 3/4, to end of
claws 19 1/2; extent of wings 31; wing from flexure 8 1/4; tail 3 3/4;
tarsus 1 4 1/2/12; middle toe 1 9 1/2/12, its claw 4/12.

In a male, the roof of the mouth is deeply concave, with a prominent

median ridge, and oblique grooves toward the end. The tongue is 1
inch 10 twelfths long, fleshy, with a deep longitudinal groove, two
lateral series of filaments, and a thin broadly rounded tip, as in other
ducks. The œsophagus, a, b, is 10 1/2 inches long, 5 twelfths in
diameter for about four inches, then enlarged to 10 twelfths, and
again contracted as it enters the thorax. The proventriculus, b b, is 1
inch and two twelfths long, its greatest diameter 8 twelfths. The
stomach, c d e, is a very large and powerful gizzard, of an elliptical
form, compressed, 1 inch and 9 twelfths long, 2 inches in its greatest
breadth, or in the direction of the lateral muscles, of which the right,
c, is 10 twelfths thick, the left, d, 9 twelfths. The epithelium is thick
and rugous; much thickened and forming two roundish, flat or slightly
concave grinding surfaces, opposite the muscles. The intestine, e f
g, is 6 feet 10 inches long, wide, its diameter for 2 feet being 4 1/2
twelfths, towards the rectum enlarging to 6 twelfths. It forms first a
very long duodenal curve, c e f g, and is then convoluted or coiled in
numerous folds. The rectum is 5 1/4 inches long; the cœca 11 inches,
their greatest diameter 6 twelfths, for 2 inches at the commencement
2 twelfths, towards the end 2 1/2 twelfths, their extremity rounded.
The trachea, h, is 7 1/2 inches long; its diameter at the upper part 4
twelfths, gradually diminishing to 3 1/2 twelfths; it then enlarges to 5
twelfths, and contracts to 3 1/2 twelfths at the commencement of the
dilatation of the inferior larynx, which is extremely similar to that of
the Widgeon, but larger; there being an enlargement, i, formed by a
number of the lower rings united, and to the left side a rounded bony
tympanum j; the greatest transverse diameter of this part, from i to j,
is 1 inch 1 twelfth. The bronchi, k k, are of moderate size, covered
with a dense layer of adipose matter.
 LEAST WATER RAIL.

Rallus jamaicensis, Gmel.

PLATE CCCXLIX. Male and Young.

My knowledge of this pretty little species is altogether derived from

Titian Peale, Esq., of Philadelphia, by whom, in October 1836, I
was favoured with the following letter:—
“I herewith send you the ‘Little Rail’ of which we were speaking
yesterday, and the letter of Dr Rowan which relates to it. The young
died soon after I received them, but the old one lived with me until
the 26th of July (four days after its capture), evincing considerable
anxiety for the young, as long as they lived. Both young and old
partook sparingly of Indian meal and water, or bread and water, and
soon became quite at home, and probably might have been
domesticated, had they been properly accommodated.
“The most remarkable part of the history of this individual is, that
after its death we should have discovered in dissection that it was a
male, rendering it singularly curious that he should have suffered
himself to be captured by hand while in defence of the young brood.
“There is now in the Museum a specimen of this species, which has
been in the collection for about thirty years, said to have been caught
in the vicinity of the city. It stands labelled ‘Little Rail, Rallus minutus,
Turton’s Linn;’ but the authenticity of the specimen has always been
disputed by Bonaparte and others, because none else had been
found; and the author just named expressed a belief that it was an
immature specimen of Rallus (Crex) Porzana of Europe.
“I regret that I should have mislaid the measurements of the
specimen when recent, if any were taken, and cannot lay my hands
on them, or any thing more than the above notes. Respectfully
yours, &c.

Titian R. Peale.”
Inclosed in Mr Peale’s letter was the following note from Dr Rowan
“to the Messrs Peales.”
“On Saturday last I wrote to you of the Rail Bird breeding near this
place. I then described one that I caught last summer, which was
unlike the Rail in the fall season, and I presumed that all in the wet
ground were the same, but this day my men mowing around the
pond started up two of the usual kind. The hen flew a few rods, and
then flew back to her young in an instant, when they caught her
together with her four young, which I herewith send you. Many more
can be caught. I have seen them in our meadow every month of the
year, but they never make a great noise except when very fat on the
wild oat’s seed. From the above you will conclude that they do not
migrate to the south, but breed here. Respectfully,

Thomas Rowan.”

Rallus jamaicensis, Brisson Sup. p. 140.—Gmel. Syst. Nat. vol. ii. p. 718.—
Lath. Ind. Ornith. vol. ii. p. 761.

Adult Male. Plate CCCXLIX. Fig. 1.

Bill shorter than the head, rather stout, compressed, tapering. Upper
mandible with the dorsal line nearly straight, being slightly convex
toward the end, the ridge narrow and convex in its whole length, the
sides convex towards the end, the edges sharp, the tip rather acute.