Spatial data and their representation

A. Arko-Adjei
Department of Geomatic Engineering
KNUST, Kumasi, Ghana
[email protected]

February 2013

1
Course content
• Introduction to GIS
• Spatial data types and representation
• Data input and methods of data capture
• Spatial referencing
• Fundamentals of remote sensing
• Sensors and platforms
• Image data characteristics and image interpretation
• Remote sensing applications

2
Lecture overview
• Geographic phenomena
• Fields
• Objects
• Computer representations
• Vector
• Raster
• GIS textbook:
• Chapter 2, Geographic information and spatial data types: Sections
2.1, 2.2 and 2.3

3
Geographic phenomena

 Defining a geographic phenomenon

• It can be named and described
• it can be geo-referenced (it has a position in space)
• it is / was present during a certain time (interval)

4
Geographic phenomena
 We distinguish between two different views of real world
phenomena: field view and object view
• Geographic fields
 For every point in a study area, a value can be
determined
 Example: elevation

• Geographic objects
 Sparsely populate a study area as well distinguishable
entities
 Example: road
5
Geographic objects
 Objects sparsely populate study area
 Space between objects is potentially empty
 Usually objects are easily distinguished and named
 Position in space is determined by one or a combination of:
• location
• shape
• size
• orientation

6
Geographic objects

 Geographic object types

• Point feature
• Line feature
• Area feature

7
Geographic objects

 Usually not considered in isolation

• As collections of objects, viewed as a unit
• Different objects do not occupy the same location

 Collection of objects can represent phenomenon at

higher aggregation level
• Forest plots form a forest
• Roads form a road network

8
Computer representations of geo-information
Important consideration:
 Complexity of reality can only be partly reflected
 In case of Fields
• Impossible to store and represent all values for all locations
• Strategy: sample of field values + mathematical function

 In case of Objects
• Object boundaries can only be approximated
• Strategy: sample of boundary locations + mathematical function

9
Computer representation of geographic features
 There are two geometric representations of geographic
features (depending on data sources and the input
method):
• Raster representation (tessellation)
• Vector representation

10
Raster representation -tessellation
 Partitioningof space into cells that together make up the
complete study area
• A thematic value is associated with each cell
• This value is associated with the entire space occupied by the cell
 Regular tessellation
• Cells have the same shape and size
 Irregular tessellation
• Cells have different shape and size

11
Raster - tessellation

 Raster,or raster map

 Raster characteristics:
• An array of m * n elements
• Raster resolution: the area each raster cell represents
• The x,y location associated with a cell is fixed by convention
e.g mid-point

 Known partition of space by raster maps allows for fast

computation (advantage)
• Cell boundaries are artificial and fixed, not adaptive to natural
phenomena (disadvantage)

12
Example of a raster cell

Raster cell,
Raster cell,
Pixel
Pixel

13
Example of geographic feature in raster form

14
Sources of raster files
 Scanned maps (requires line following processing and/or
raster-to-vector conversion)
 Existing digital raster data
 Remote sensing data e.g. air photos, satellite imagery
requires classification into mapping categories
Vector representation
 Tryingto explicitly associate co-ordinate pairs with
geographic phenomena

y
Farm 1

Farm 2

16
 Vector representation

 Point representations
 Line representations
 Area representations

A small-scale house A small-scale road A large-scale house

Point representation Line representation Polygon representation

17
Vector representation
• Point representation
• used to represent point locations
• e.g. addresses, elevation spot heights, locations of malls, banks,
cities, volcanoes, etc.

A small-scale house
Vector representation
• Line feature
• consists of an ordered set of connected points
• e.g. political boundary, contours, geological faults, streets,
highways, rivers, etc.

A small-scale road
Vector representation
• Area feature
• used to represent a region
• e.g. forests stand, climatic zones, lakes, soil types, land use,
nations, counties, etc.

A large-scale house
Vector vs raster representations

V e c to r d a ta m o d e l
y y y code

code

code

x x x
P o in t L in e A re a
R a s te r d a ta m o d e l
Row

Row
Row

C o lu m n C o lu m n C o lu m n
K e y f o r t h e v e c t o r m o d e ls : in t e m e d ia t e p o in t
node

21
Raster and Vector:
Representations of fields and objects

Raster model Vector model

Simple data structure Complex data structure

Easy and efficient overlaying Difficult to perform overlaying
Compatible with Remote Sensing imagery Not compatible with RS imagery
High spatial variability is efficiently represented Inefficient representation of high spatial variability
Simple for programming by user
Same grid cell definition for various attributes

Inefficient use of computer storage Compact data structure

Errors in perimeter and shape Efficient encoding of topology
Difficult to perform network analysis Easy to perform network analysis
Inefficient projection transformations
Loss of information when using large pixel sizes
Less accurate and less appealing map output Highly accurate map output

22
 Layers
 Even simple spatial applications have multiple sources for
their geographic data.
 Itis not considered good practice to store all of these in the
same subdivision of space.
 Data whether raster or vector results in many possible
subdivisions of space, simply depending on the point of view
we take.
Layers

 A better approach is to represent each of these separately, in

what is known as a spatial data layer.
 Layers are spatially related map-like subdivisions of space,
usually based on a single perspective.
 Each layer contains information about one class of spatial
phenomena .
Layers

Roads

Bus route

Shopping centre

Industrial sites

Lat
itu Longitude
de
Topology and spatial relationships

Topology
 A branch of mathematics that deals with properties of
space that do not change under certain transformation
 Dealing with spatial relationships

 ‘Rubber sheet’ transformation

• shape and size of properties may change,
• definition of and spatial relationships between properties are
maintained

26
Topology and spatial relationships
 Rubber sheet transformation

1
1
C 5
C 5 A
A
D E7 D E 7
6 6
4 3 4 3
B B
2 2

27
Topology and spatial relationships
 Topological invariants

Interior Boundary
Exterior

28
Topology and spatial relationships

Topological relationships
 Relationships between two regions can be determined based
on the intersection of their boundaries and interiors (4-
intersection).

A B

29
Topology and spatial relationships

disjoint covered by

meet contains

equal covers

inside overlap

30
Topology and spatial relationships in summary:
 The main elements of topology are
• topological consistency relations
• relationships between nodes, lines and polygons

Spatial relationships
E (examples)
c G A inside B
D meets B
B A C disjoint from B
G overlaps E
F
D

31
Geographic information and spatial data types
In conclusion:
 Computer representations
• Raster representation
• Vector representation

 Geographic phenomena
• Fields
• Objects

32