Information
The 
raster
model of space is more suited to representing continuously varying phenomena such as land cover,
elevation, and in our example, population density. Unlike the points, lines and polygons we have considered previously, it
does not require
us to define artificial boundaries around individual features.
Google Earth
makes significant use of rasters in the form of satellite and aerial photography. But in the
GI science
context, a raster is more than simply a bitmapped image. It is located in space, usually by
defining an origin using some georeferenced coordinate system (e.g. longitude/latitude coordinate pair). The raster is composed
of
cells
of a fixed resolution
that together determine the level of detail
contained within the raster and its spatial extent. Unlike a bitmapped image, the individual raster cells can represent more
than simply a
colour. The cell's value can represent any measurable or countable phenomenon, such as elevation, population density or temperature.
It can
use non-integer numbers as well as null values to represent absent data. A separate colour lookup table that maps measurement
or count to
colour value is then used to provide a visual representation of these numeric values. As we have seen, cartographic research
has resulted
in suggestions for a number of effective colour schemes that relate well to particular kinds of variation in the data and
means of display
(Harrower et al. 2003).
Google Earth
can display rasters using its base aerial photography and through imported images
georeferenced with GroundOverlay (see example #5 in section 1.1.5). However, as is the case with other
geobrowsers
it currently has very limited functionality for processing rasters. Therefore in order to
enhance the geovisualization process with raster analysis, we must rely on other software to perform the necessary raster
generation and
processing.
Example
We will use LandSerf
to perform the raster analysis of our population and photographic data in this tutorial.
LandSerf
contains much sophisticated analytical functionality that we will not be touching on here.
In this tutorial, we will simply consider some of the raster and vector display options available in LandSerf and show how
these may
contribute some ideas from GI science
to a geobrowser mashup.
LandSerf Graphical User Interface showing thumbnail and main views of raster and vector data.
The LandSerf
GIS
is designed to be a primarily visual interface to
geographic information and analysis. It allows multiple raster and vector maps to be stored and processed. A list of all stored
maps is shown
as a series of thumbnails down the left-hand side of the main window. Maps can be classified as either primary
or secondary. Primary maps are indicated as blue shaded thumbnails and secondary maps as pink thumbnails.
Clicking on a thumbnail with the left mouse button selects it as a primary raster or vector map. Clicking with the right mouse
button selects
it as a secondary map (users of single-button mice can shift-click to access right-click functionality). The primary raster
and vector maps can
be displayed in the main LandSerf
window, which can be resized at any time. To change the maps displayed in the
main window, click on the appropriate thumbnail and select the menu item or the circular
arrows button.
To load a map into LandSerf
, use the menu item and
select a file in either '.srf' (raster) or '.vec' (vector) format. Raster maps are displayed automatically once they are opened.
To toggle
vector map display on or off, select the menu item. Once displayed, you can zoom
and pan around the map by toggling the menu item or the magnifying glass button.
Dragging with the left mouse button zooms in and out of the map while dragging with the right pans across it. If you are using
a one-button
mouse, shift-dragging can be used to pan across the map. The menu item or
double headed arrow button will return to the full map display by fitting the maps into the current window.
Raster and vector values can be interrogated by toggling the menu item, or the
question-mark button. When selected, moving the mouse over a raster will display its geographic location and the attribute
of the primary raster
at the mouse location, in the status bar at the bottom of the LandSerf
window. Clicking on the mouse will display
the secondary raster and vector attributes at that location.
Finally, the appearance of the vector maps can be controlled by selecting the
menu item. Of most use here are likely to be the ,
and options.
LandSerf display showing shaded relief representation of the population density raster with vector state population values
overlaid. Line width is set to 0.5 pixels and polygon opacity low to allow underlying shaded
relief to remain visible.
