Representing shape
Geographic information systems allow us to represent,
investigate and analyse the real world using a computer. A computer,
at the fundamental
level, is little
more than millions of switches processing electronic signals. To work, a
computer relies on the things that it is processing being reduced to
a very simple level.
Therefore, systems that work with geographic information need to break the
geographic information down to very simplified versions of reality so the
GIS can work with them.
There are two methods by which geographic information is broken down for
use in a computer:
1. Physical features and human features are reduced to points, lines and
polygons which are linked to information that describes them (their attributes).
This
method of representing the world around us is called vector
2. Physical features and human features are divided into a regular grid with
each cell in the grid holding descriptive information about its contents.
This method of representing the world around us is called raster.
If you wish to use a geographic information system it is important to understand
these two methods of representation:
Representing with vector
The key to understanding the vector method is realizing that all shapes,
no matter how complicated, can be simplified to a number of very simple tables
containing information about the shapes that are linked together. The
vector method of representing the real world breaks down the shape component
of the geographic information into simple tables of information that a computer
can easily work with. It then links the tables together using a database to
draw the shapes.
The key to understanding how a geographic information system works with the
shape component of geographic information is to understand how complex shapes
can be reduced to linked tables of information:
Points
Points are the simplest shape that can be associated with geographic information.
They have zero dimensions and can therefore be located using a simple, single
x,y coordinate located relevant to a coordinate system. Points can represent
physical features (eg. a telegraph pole) or non-physical features (eg. the
location of a road accident). Storing point locations is very simple and requires
nothing more than a single table containing columns for the x and y coordinate
locations and ID numbers, one for each point:
Lines
Lines are an extremely common form of one-dimensional geographic information.
They are used to represent physical features such as roads and railways and
contour lines.
Storing lines is more complex than storing points because lines change shape
when distorted. We already know that distortion is an essential aspect of representing
geographic information (eg. via map projections), and it therefore follows
that the shape of a line in one map projection might be very different in another.
Consequently, geographical information composed of lines cannot be stored by
referring to the line shape alone.
To get around this problem, lines are stored using the properties that remain
invariant of distortion (known as topology). There are only three of these:
-
Start points
-
End points
-
Intersections
All lines must have a start and an end point, irrespective of distortion.
Also, two lines that intersect will always intersect no matter how much
they are distorted.
The locations of these topological properties (called nodes) can be stored
as individual coordinate pairs. The nodes can then be connected together with
lines (called arcs) composed of ordered coordinate locations that define the
line shape (including the locations of the nodes). Storing lines in this way
requires a large number of tables of information that are linked together.
This, in turn, means that a system for handling large numbers of linked tables
of information is required. Commonly such as system is referred to as a database
and it common to find a database at the center of large datasets of geographical
information.
The following example shows how geographic information composed of lines is
stored using topology tables. The use of topology means that what appears to
be two lines is actually stored as four.
Polygons
polygons enclose two-dimensional areas. Any geographic information relating
to enclosed boundaries or areas is represented using polygons. The sorts of
features that polygons represent include the outlines of lakes and reservoirs,
the outlines of buildings, country outlines and political boundaries.
topology is also used in storing the coordinates that represent polygons because
polygons are fundamentally composed of arcs. Below the two polygons (Poly A
and Poly B) are composed of three arcs which, in turn, are composed of two
nodes and a number of coordinate pairs. As before, the shapes can be reduced
to a number of linked tables (a database) that, together, hold all of the information
necessary to draw the shapes in a map space.
Using topology to build polygons also has the enormous advantage that lines
are not doubled up at the point at which two adjacent polygons juxtapose. As
for lines, a large number of tables of data are required and so a database
is used to organize and manage the data.
Representing with raster
Physical features and human features can usually be divided into a grid and
this is a much more simple way of representing the world around us than using
points, lines and polygons.
Take the following example of a lake in a park that has been represented by
dividing the geographic information into a grid
This is the process of representing the world using raster. raster is an extremely
simple method of representing geographic information and, therefore, the
computer in a geographic information system is able to work with raster very
easily. raster does not require the extensive database of linked tables to
store the shapes of the features.
The main drawback of raster lies in its representation. Whilst many physical
features and human features naturally form points, lines and polygons, very
few naturally form a grid pattern. Therefore, raster is limited in the reality
with which it can represent geographic information and the detail that the
raster data is able to provide is clearly limited by the size of the grid that
is used.