density data sets were 
The populational 
visualized and manipulated with 
Sao Paulo State Map with municipalities boundary 
(IGC, 1984); 
the software SGI, 
Firstly, the boundary regions were digitilized and 
population densities by government regions about 
stored with vector structure. 
1980, 1985 and 1991 (SEADE, 1992). 
developed at the National Institute for Space Research - 
INPE (Brazil), to verify the viability and efficiency of this 
was developed to run in IBM-PC or compatibles, with 
kind of Cartographic Information representation. The SGI 
operation system DOS. Data input in SGI is possible by 
In this case, each polygon 
digitizing or importing files with DXF or ASCII formats. In 
this software, the DTM is represented by a rectangular 
grid or triangular irregular grid (TIN). 
represents a government region. Afterwards, this vector 
structure was transformed in a raster structure, and the 
software did generate a rectangular grid making the 
correspondence between each pixel 
and population 
density value of government regions. This rectangular 
grid represents the 3D surface of choropleth mapping. 
2. GENERATING DIGITAL MODEL FOR CHOROPLETH 
In the SGI, a DTM can be represented either by 
a set of contour lines, by a gray level map, or by means of 
a 3D visualization, among other possibilities. The 3D 
visualization can be defined in perspective or parallel 
MAPPING 
In generating digital models for choropleth 
mapping, the value in the enumeration unit must be 
considered uniformly spread throughout the unit. The top 
projection. To population density values about 1980, 
1985 and 1991 
digital model 
of each prism in the digital model is horizontal and 
unchanging. Therefore, in the rectangular grid, all points 
3 grids were obtained. In the figure 1, the 
of 1991 may be observed 
in a 3D 
that correspond a specific area must have the same value 
visualization, using parallel projection. 
Z. In SGI, the first step to obtain this grid, is to store the 
polygons that correspond the enumeration units. The 
available data, to develop this work, were: 
Figure 1 - Digital model of population density of 1991 
    
  
  
              
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
           
697 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996 
  
on 
ay 
d, 
us 
ne 
ta 
re 
ne 
CS 
in 
de 
de 
es 
es 
es 
lo, 
es 
its 
he 
to 
he 
of 
Ire 
re 
ilo 
Ire 
es 
43 
he 
he 
1S. 
"S 
nd 
00