VISUALIZATION OF CARTOGRAPHIC INFORMATION USING DTM 
Claudia Robbi 
National Institute for Space Research - INPE 
Av. dos Astronautas, 1758 
12227-010 Säo Jose dos Campos - SP, Brazil 
e-mail: robbi@ltid.inpe.br 
KEY WORDS: Cartography, DTM, Thematic, Mapping, Three-dimensional visualization, statistical information 
visualization 
ABSTRACT 
The representation of geographical phenomena by two-dimensional maps presents serious limitations that may 
be overcome by means of a three-dimensional (3D) model. The advent of powerfull enough computers, and, 
consequently, Geographical Information Systems (GIS) allows the creation of models which represent continuous 
spacial phenomena, and its visualization by Computer Graphics. Softwares that have been developed to represent the 
relief of the Earth’s surface through digital terrain models (DTM’s) could be used to visualize many other different data 
sets. In this work, 3D models are generated from population density data sets of Sao Paulo State, Brazil, which were 
obtained from Annual Report of Sao Paulo State (SEADE, 1992). Choropleth maps were produced through the 
classification of the DTM from data variation, called “DTM slicing”. With “DTM slicing”, analyses of data caracteristics 
are possible, such as “what are regions that have more than 500 inhabitants per km”, or regions with reduction in 
population? 
RESUME 
La représentation de phénoménes géographiques au moyen de cartes bi-dimentionelles présentent de 
sérieuses difficultés qui peuvent être contournées au moyen de l'utilisation de modèles tridimentionels. L'apparition de 
calculateurs suffisamment puissants et, en conséquence, de Systèmes d'Information Géographiques (SIG), permet la 
création de modèles qui représentent des phénomènes spaciaux continus, et leur visualisation par infographie. Les 
logiciels qui ont été développés pour représenter le relief de la superficie terrestre, à travers des modèles numériques 
d'élévation (MNE), peuvent être utilisés pour visualiser bien d'autres ensembles de données. Dans ce travail, les 
modèles 3D ont été obtenus à partir d'ensembles de données sur la densité de population dans l'état de Säo Paulo, 
Brésil, qui ont été extraits du Rapport Annuel de l'État de Säo Paulo. Des cartes choropléthes ont été produites à 
travers une classification du MNE, appelé "découpe en tranche du MNE". Au moyen de cette "découpe", des analyses 
sur les caractéristiques des données sont possibles, telles que "quelles sont les régions qui ont plus de 500 habitants 
au km2, ou quelles sont les régions qui ont eu une diminution de population?" 
1. INTRODUCTION statistical surfaces, through a regular rectangular grid, the 
operations with these grids permit the user of GIS to 
The statistical data sets aggregated by statistical acquire and to observe different information about the 
or administrative areas can be represented through phenomena, like population distribution. One kind of 
choropleth maps. The choropleth mapping may be operation is to classify data variation, with a procedure 
thought of as a three-dimensional (3D) statistical surface. called “DTM slicing”. 
This surface is composed by prism. The height of each 
prism is proportional to the value it represents, and the In this work, three-dimensional models are 
bases are defined by the frontiers of statistical or generated from population density sets of Sao Paulo 
administrative areas (DENT, 1986). The visualization of State, Brazil (SEADE, 1992). The data sets are 
these data models sometimes are called prism maps. aggregated by government regions that encompasses 
municipalities. Sao Paulo State has 624 municipalities, 43 
The digital models, that represent surfaces, are government regions and 15 administrative regions. The 
obtained by interpolation techniques. If the geographical municipalities form the government regions, and the 
location and attribute Z, of the points set, are known, a government constitute regions the administrative regions. 
regular rectangular grid can be generated. The 'Z' value Sao Paulo State is situated between latitudes 19°47'22"S 
of grid points are calculated by interpolation. The and 25?18'35"S, and longitudes 44°09'46"W and 
statistical information represented with choropleth maps 53°05'15"W. The state has approximately 31000000 
have its geographical location known, and the statistical inhabitants, according the 1991 census (SEADE, 1992). 
information is the attribute itself, or 'Z' value of the digital 
model. If a digital model represents appropriately these 
696 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996 
  
HS D PR pe a une na 5" ih 
AY "rt Ni] ee $ gn o oy