As you recall, we discussed four general classes of raster-based
GIS operations: reclassification, overlay, distance,
and neighborhood characterization. We have already examined
aspects of the former two in previous exercises. We now turn our
attention to applications of the latter two. Distance
operators can be Euclidean (as the crow flies) or
weighted (non-Euclidean or as the intoxicated
person walks). Also, we can think of distance not only in
units of feet, meters, miles, kilometers, and so on, but also
in terms of effective time or accumulated cost.
Context operators examine neighborhoods of
grid cells and summarize their values based on some mathematical
or statistical function, such as minimum, maximum, mean, median,
and so on.
In Idrisi there are several Distance-
and Context-related operations. In this exercise, we will examine
and apply two of those, respectively: DISTANCE and SURFACE. The
former measures the Euclidean distance between each cell and the
nearest of a set of target features. Distances are output in reference
system units. Distances derived from images with rectangular cells
are correctly calculated. DISTANCE first requires that you input
the name of the feature image. This is the image containing the
target features from which distance should be measured. This image
must contain integer values with non-zero cells indicating the
feature targets. Then input a new name for the output image, as
well as the value units and a title. Related to DISTANCE is COST,
which generates a distance/proximity surface (also referred to
as a cost surface) where distance is measured as the least effort
in moving over a friction surface. The unit of measurement is
grid cell equivalents (gce). A grid cell equivalent of
1 indicates the cost of moving through a grid cell when the friction
equals 1. A cost of 5 gce's might arise from a movement through
5 cells with a friction of 1, or 1 cell with a friction of 5.
Costs are determined radially from a set of source targets to
the edges of the image. We will deal with cost surfaces in a subsequent
A context operator, SURFACE calculates slope, aspect and shaded
relief images from a digital elevation model. SURFACE requires
that you first choose whether slope, aspect, slope and aspect,
or analytical hillshading should be calculated. You are then required
to input the name of the digital elevation model image and a name(s)
for the output file(s). Choose whether you want the calculations
in degrees or percent and enter the conversion factor (see Note
4). The default is set to 1. If you have chosen to calculate analytical
hillshading, the calculations are always in degrees and you must
input both the sun azimuth in degrees and the sun elevation angle
in degrees. The defaults are set to 315 and 30 respectively. Finally,
enter a title(s) for the output image(s).
SURFACE determines the slope by calculating the maximum slope
around each pixel from the local slopes in X and Y. Only the neighbors
above, below, and to either side of the pixel are accounted for
in this procedure, called a "rook's case" procedure.
Slopes are output in either decimal degrees or percents. Slope
gradients in percents represent the tangent of the angle multiplied
by 100. Therefore, a 45 degree angle = 100%, while a 90 degree
angle will approach an infinite percent.
The aspect is the direction in which the maximum slope faces.
Aspects are output in decimal degrees and use standard azimuth
designations, 0 - 360, clockwise from north. In regions where
the surface is perfectly flat with slope = 0, aspect is assigned
a value of -1. The output data type is real.
Analytically shaded images are for purposes of illustration only
and are best viewed with the Grey 16 palette. The numeric results
represent only relative illumination and are not quantified to
any external reference. The resulting image may benefit from a
If an image's value units do not match the reference system units,
SURFACE requires a conversion factor to calculate the slope. The
conversion factor is the number with which you multiply values
in one unit (e.g., the value units) to change them to the values
in the other unit (e.g., the reference units). Check the documentation
file to confirm the units before proceeding. For example, if the
value units are "ft", but the reference units are "m",
the conversion factor would be 0.3048; a conversion in the opposite
direction (from "m" to "ft") would require
a conversion factor of 3.28. Conversion in the case of "deg"
or "rad" are more complicated, because the ground distance
between degrees changes with latitude.
To perform the algorithms on the edge values, SURFACE temporarily
adds an extra row and column around the edges of the image and
assigns the cells the values of the adjacent cells.
This exercise, as described in the Idrisi
for Windows Student Manual, is a suitability assessment problem
in which we are to locate all areas meeting siting criteria and
constraints for a light manufacturing plant. Site suitability
criteria include acceptable terrain slope (< 2.5 degrees),
size (contiguous suitable areas of at least 10 hectares), current
land use (must be within currently forested area), and water protection
(sites need to be more than 250 meters from reservoirs).
To solve this problem, you will need to use both DISTANCE and
SURFACE, as well as some of the operators you have used previously,
such as RECLASS, ASSIGN, OVERLAY, GROUP, AREA, and a couple others.
This exercise is an excellent one for getting a real feel for
cartographic modeling, or digital light table gymnastics. You
might wish to start the exercise by overviewing the model flowchart
With the supervision of your instructor or lab coordinator, use the Idrisi for Windows software to carry-out the directions for Tutorial 8 in the Student Manual.
The data for this tutorial are located in the Idrisi
for Windows sample exercise data directory, typically C:\EXERCISE.
The path may vary from PC to PC, so check with the instructor
or lab coordinator.
For Tutorial 8 on Distance and Context Operators, answer
questions 4, 5, and 6.
For map results, produce a hardcopy of SUITABLE, which is the
final result of the cartographic model, as well as a hardcopy
of an orthographic projection of a version of this suitability
map draped over the terrain model named RELIEF. As before, please
produce an electronic copy of each of these maps, using whatever
graphics tools you prefer, in either a .BMP or .GIF format.
For University of Hartford students, submit your answers and digital maps (images) via e-mail to:
For students at UConn, submit your work to:
If time permits, explore the ramifications of changing one or
more of the siting criteria, varying only one criterion at a time
(initially, anyway). For example, you might consider relaxing
the slope constraint to allow the facility to be located on slopes
up to 4 degrees. Or maybe allowing it to be located within 150
meters of reservoirs. Or maybe pushing it back to no closer than
350 meters. Perhaps the manufacturing plant can be situated on
parcels that are only 5 hectares in extent. There are many variations
one could implement in this if-then-what type of
modeling. That's the real power of computer-based cartographic