Vector Data Model

Report
GI Systems and Science
January 23, 2012
Points to Cover
What is spatial data modeling?
 Entity definition
 Topology
 Spatial data models

 Raster data model
 Vector data model

Representing surfaces using
 Raster approach
 Vector approach
Spatial Data Modeling
GIS are computer representations of the real
world
 These representations are necessarily simplified

 Only those aspects that are deemed important are
included

The simplified representation of the real world
adopted by GIS is a model
 Set of rules about how the spatial objects and
relationships between them should be represented
Spatial Data Modeling

A GIS model can be conceptualized in terms of
two aspects:
 A model of spatial form: how geographical features
are represented
 A model of spatial processes: how relationships
between these features are represented

Building a model of the world for your GIS is a
key stage in any GIS project
Formulating
research
question
Collecting
data
Creating
data model
Entering
data into a
GIS
Spatial Data Modeling

Creating a data model involves going through a
series of stages of data abstraction:
 Indentifying the spatial features form the real world
that are of interest in the context of the research
question
 Choosing how to represent the features (i.e., as
points, lines or areas)
 Choosing an appropriate spatial data model (i.e.,
raster or vector)
 Selecting an appropriate spatial data structure to
store the model within the computer
Formulating
research
question
Collecting
data
Creating
data model
Entering
data into a
GIS
Entity Definition
Figure 3.2
Source: Heywood et al., 2011
Entity Definition

Surfaces
 used to represent continuous features or phenomena
Figure 3.3
Source: Heywood et al., 2011
Entity Definition

Networks
 used to represent a series of interconnected lines
along which there a flow of data, objects or materials
Figure 3.5
Source: Heywood et al., 2011
Entity Definition

Issues associated with simplifying the
complexities of the real world
 Identification of the proper scale for
representation
 How much detail is required?
 Dynamic nature of the real world
 How to select the most appropriate representation of
the feature?
 How to model change over time?
 Identification of discrete and continuous features
 Fuzzy boundaries
Entity Definition

Features with
fuzzy boundaries
 Continuous canopy
and open woodland
Figure 3.7
Source: Heywood et al., 2011
Topology

A geometric relationship between objects
located in space
 Adjacency
 Features share a common boundary
 Containment
 A feature is completely located within another feature
 Connectivity
 A features is linked to another feature
Independent of a coordinate system
 Independent of scale

Spatial Data Modeling

Creating a data model involves going through a
series of stages of data abstraction:
 Indentifying the spatial features form the real world
that are of interest in the context of the research
question
 Choosing how to represent the features (i.e., as
points, lines or areas)
 Choosing an appropriate spatial data model (i.e.,
raster or vector)
 Selecting an appropriate spatial data structure to
store the model within the computer
Formulating
research
question
Collecting
data
Creating
data model
Entering
data into a
GIS
Spatial Data Models
Data models and corresponding data structures
provide the information the computer requires to
construct the spatial data model in digital form
 Two main ways in which computers can handle
and display spatial entities:

 Raster approach
 Vector approach
Spatial Data
Models

The raster data model
 Based on principles of
tessellation
 Cells are used as
building blocks to create
images of features
 The size of the cell
defines the resolution
(degree of precision)
with which entities are
represented
Figure 3.8
Source: Heywood et al., 2011
Spatial Data
Models

The vector data model
 The real world is
represented using twodimensional Cartesian
co-ordinate space
 Points are basic building
blocks
 The more complex the
shape of a feature the
greater number of
points is required to
represent it
Figure 3.8
Source: Heywood et al., 2011
Raster Data Model

Basic raster data structure
 One layer stores and represents one feature
 Presence-absence principle
Figure 3.10
Source: Heywood et al., 2011
Raster Data Model

Raster file structure for storing data on several
entities of the same type
Figure 3.11
Source: Heywood et al., 2011
Raster Data Model

One of the major problems with raster datasets is
their size
 A value must be recorded and stored for each cell in an
image regardless of the complexity of the image

To address this problem a range of data
compaction methods have been developed
 Run length encoding
 Block coding
 Chain coding
 Quadtree data structures
Raster Data Model

Raster structure for storing data on several entities of
the same type
 Reduces data volume on a row by row basis
Figure 3.12(a)
Source: Heywood et al., 2011
Vector Data Model

Basic vector data structure
 A file containing (x,y) co-ordinate pairs that represent
the location of individual points
Figure 3.14(a)
Source: Heywood et al., 2011
Vector Data Model

Point dictionary vector
data structure
 Allows to avoid
redundancy when areal
features share a
boundary (are adjacent)
 But does not really
store information on
topology
Figure 3.14(b)
Source: Heywood et al., 2011
Vector Data Model

Topological vector
data structure
 Informs the computer
where one feature is in
respect to its
neighbours
 Withstands
transformations well
Figure 3.15
Source: Heywood et al., 2011
Vector Data Model

All topological vector data structures are
designed to ensure that:
 Nodes and lines segments (arcs) are not duplicated
 Arcs and nodes can be referenced to more than one
polygon
 All polygons have unique identifiers
 Island and hole polygons can be adequately
represented
Modeling Surfaces

Surfaces represent continuous features of
phenomena
 Theoretically have an infinite number of data points

A model of a surface approximates continuous
surface using a finite number of observations
 The issue of selecting a sufficient number
observations
Modeling Surfaces

Digital Terrain Models (DTMs) are digital
datasets recreating topographic surfaces
 Created from a series of (x,y,z) data points
 Resolution is determined by the frequency of
observations used
 Are derived from a number of data sources
 Maps (low to moderate accuracy, all scales, selected
coverage)
 GPS (high accuracy, small areas)
 Aerial photographs (high accuracy, large areas)
Modeling
Surfaces

Raster approach
 DTM is a grid of height
values
 Also known as Digital Elevation Model
(DEM)
 Each cell contains a value representing the
height of the terrain covered by the cell
 Accuracy depends on the size of the cell and
complexity of the surface
Figure 3.21
Source: Heywood et al., 2011
Modeling Surfaces

Vector approach
 Grid
 Triangulated Irregular
Network (TIN)
○ Triangles provide
area, gradient and
aspect of terrain
○ TINs use only surface
significant points to
reproduce a terrain
surface
Figure 3.22
Source: Heywood et al., 2011

similar documents