Once geodetic control data were obtained from previous surveys of the 
area, a field check was conducted to locate and photo identify the 
available survey markers. A block triangulation and error propogation 
was then performed on the complete fourteen model block of temporal 
images. The unit variance of the final block adjustment was ./91 with 
661 degrees of freedom. 
The project area boundaries were defined and geounits (project sub- 
areas) were calculated to coincide with the actual areas contained in 
each of the stereo models. The geounits could then be completely digi- 
tized without constantly changing photos on the analytical plotter and 
unnecessarily performing rigorous set-up and registration procedures. 
Classification files were also created to contain the specific attribute 
information. 
Data collection began through on-line photo interpretation and manual 
digitization using the CAPIR system. Three-dimensional terrain informa- 
tion was collected theme-by-theme and geounit-by-geounit. The road net- 
work layer was collected first by digitizing the road center and assign- 
ing surface material and road width attributes. Surface drainage was 
delineated in the same manner by digitizing the center of streams and 
gullies. These two layers consist solely of linear features. There- 
fore, only the center attribute was entered leaving the left and right 
attributes of the line being digitized empty or null. Land cover was 
defined as the terrain surface covered by both natural vegetation and 
man-made polygonal features. Areal information of this type were 
entered with attributes for each of the left, center, right portions of 
the line being digitized for defining polygons. 
The digital elevation data (DEM) are being collected with the CAPIR/DEM 
profiling system on the analytical plotter. This is a collection system 
which uses an adaptive spacing grid to optimize profiling in areas which 
exhibit significant topographic disparity (Edwards et. al., 1986). For 
instance, areas of different sizes, corresponding to individual 
geomorphic features or other regions, may be sampled at a wide range of 
posting densities. The horizontal resolution of the data has a 5 meter 
spacing in those areas which may be traversed by the ALV. Elevation 
data for rugged areas will be collected at a higher spacing, interpo- 
lated to 5 meters, and flagged within the DEM file. The soil theme con- 
sists of photo-interpreted soil materials. Emphasis will be placed on 
mapping materials that would hinder mobility when wet. The landform 
layer will consist of photo-interpreted geomorphologic structures. This 
information will be useful as a qualitative check on other thematic 
layers and could be used in resolving uncertainty. 
Once data capture was completed, topological verification was performed 
for each geounit in every layer. When verification was accomplished the 
data was reformatted from an arc/node structure into a polygonal struc- 
ture to be compatible with other system components. 
481