Figure 5 - Population Density 1980 (inhab. per km?) - Natural breaks classification 
  
acima de 1500 
In the development of this work, it can observed 
that the utilization of DTM's to visualize and classify 
statistical information has some advantages. In the DTM 
the unclassed information are stored, and the user can 
access the data, without losing details that are 
generalized and simplified when the data are classified. 
The DTM is more useful to user, instead of a unclassed 
choropleth map, because the user can generate any 
classification map that he/she needs, without losing the 
original data. In this way, the user doesn't need 
interpreting the  unclassed information, he/she will 
interpret the results of the classification that he/she 
automatically generated. 
The result of the classification of statistical 
information, using DTM slicing in this digital model, is 
displayed in the screen, by a choropleth map, because 
the digital model stores constant values in the 
enumeration areas. When the DTM slicing is developed, 
the result contours of the classification DTM are the 
boundary areas itself. This fact permits to eliminate the 
necessary work when the data are classified separately, 
and then each class is associated with the correspond 
enumeration unit to generate the map. Even with 
automatic procedure, the DTM slicing avoids this search. 
If the boundary enumeration units is stored in a 
vector structure, it is possible overlap the classified image 
with the boundary map. This permits the user to identify 
the units that correspond represented classed, as shown 
in figures 3, 4 and 5. 
700 
4. ARITHMETIC OPERATIONS WITH RECTANGULAR 
GRIDS 
In the SGI, it is possible to realize arithmetic 
operations with two rectangular grids. Among this 
operations are: subtraction, addition, multiplication, 
division, mean. The subtraction operation between two 
grids is advantageous for representation of statistical data 
with DTM. If the user has statistical data of two different 
periods, he/she can generate two grids, for each period, 
and subtracting the grids, analyse the variation of 
phenomena in that period. 
A exemple that presents the population growth, 
demonstrates this possibility to analyse statistical data. 
This exemple about the period between 1980 and 1991 is 
shown in figure 6. The procedure consists of: 
e generating the two grids; 
e subtrating from the 1991's grid the1980's grid values. 
The resultant grid represents the population growth in 
the period; 
e visualizing this grids in 3D; 
e slicing this DTM to observe the classes of population 
growth. 
This classes can be used to specific analyses 
the user needs to develop. Figure 6 presents the 
choropleth map that represents the population growth in 
the periods mencioned above. This operation permit, for 
exemple, to observe that there was populational reduction 
in Adamantina region, between -3 to O inhab. per km”, 
and high population growth in Säo Paulo and Campinas 
regions, between 200 to 350 inhab. per km. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996 
  
* — wpa 09 rM RAT Lo AY OY 
Ff + 
Ld 
“ (5D 03 CA