P2-5-1 
Generation of 3D View Map Using by Raster Base Data Processing 
Kunihiko Ono*, Shunji Murai '**, Vivarad Phonekeo ** and Shigetaka Yasue* 
* Geoinfo Creative Co.Ltd., 2-4 Kanda Jinbocho, Chiyoda-ku, Tokyo 101, Japan. 
**Space Technology Applications and Research Program (STAR), Asian Institute of Technology, 
P.O.Box 4, Khlong Luang, Pathumthani 12120, Thailand 
KEY WORDS: 3D View Map, Polygon Shift Method, Shadow Analysis, Hidden Point Algorithm. 
ABSTRACT 
The authors have developed a new method called “Polygon Shift Method” (named so by the authors) that enables to 
the generation of a 3D view map of a city with tall buildings with a simplified procedure to shift a polygon and check 
the overlap between the original and shifted polygon. Boolean operations are applied with a newly defined “Fore or 
Aft” side and a “Depth Distance” that functionize the visibility criteria or hidden point processing in the 3D view. 
Because the polygon shift method can be operated with raster based structure the computer processing was effective in 
visualizing a 3D view map of buildings with shadow. 
1. INTRODUCTION 
Perspective of a building or landscape used to be 
drawn manually by architects. Since computer aided 
design (CAD) was introduced in the 1970’s, automated 
drawing became popular (Murai, 1997,Wilson 1998, 
etc.). In the recent decade, virtual reality software is 
available to produce perspectives or bird’s eye view of 
three dimensional structures. 
However, data format and structure are not yet very 
simple, because all details not only of the building plan 
but also nodes, edges and surfaces (vertical walls and 
roofs) should be input with specified topology. 
Various data structures of 3D GIS has been proposed 
by Molenaar (1990), Shibasaki (1992), Chen (1994), 
Temfli (1997), William (1997) etc. All these data 
structures are based on “vector base” topology. 
The polygon shift method proposed in this paper needs 
a simple data structure of a polygon (building plan 
given by vector data as the input data) to generate a 
raster belonging to the polygon. Data about roof and 
vertical walls are not necessary because the method is 
only targeted to column shaped buildings with flat and 
horizontal roofs though any complicated buildings can 
be formed with a combination of multiple columns. 
Using vector data, the geometry to achieve hidden 
point/line/surface processing for multiple buildings 
needs a very complicated algorithm when the building 
shape includes concave parts. For example, the 
intersection of lines and surfaces and the identification 
of overlapped areas of two polygons will be a 
complicated calculation for concave polygons. A 
comparison between the proposed method and existing 
algorithms of Z-Buffer, Scan-Line and Back-Face 
Removal is described in the section 5. 
In this study, a raster mode approach with Boolean 
operations works effectively to identify overlapping 
areas using the depth distance, which is a criterion to 
judge which part of a building hides other buildings. 
2. PROPOSED 3D COORDINATE SYSTEM. 
In the conventional method, most of the 3D view maps 
used to be based on xyz with a horizontal x-axis, an 
oblique y-axis and a vertical axis as shown in Fig. I (a). 
In this study, we use a different 3D coordinate system 
called Military Projection with an orthogonal xy 
system and an oblique z-axis as shown in Fig. 1(b). 
The benefit of this system is that we don’t need to 
change the original shape of its building plan, while 
the disadvantage is that the view map looks a little bit 
strange with oblique buildings. 
1 Corresponding author: E-mail: murai@ait.ac.th