Masafumi Uehara 
  
We resolve the above problem by utilizing a large scale digital map. In the last few years, a digital map has been 
spreading remarkably. The digital map we used is expressed the topographical shape of buildings accurate 
comparatively (see Figure 2), so we know where buildings are separated. And, considered utilization of wide area city 
models, it is necessary to prepare data based on a digital map for modeling. Therefore, in this paper, we propose a 
method of the construction of urban scene model using a digital map adding 3D data of buildings by motion image 
analysis. 
First, we mention the method how to obtain 3D data of buildings in Chapter 2, secondly Chapter 3 describes the 
matching method with measured 3D data and a digital map data. Experimental results are shown in Chapter 4, and 
Chapter 6 describes conclusion of this research. 
2 IMAGE SEQUENCE ANALYSIS 
. 2.1 Obtaining depth data by EPI analysis 
3D data of buildings are obtained by EPI analysis. The 
traces of feature points appear on the Epipolar plane 
  
   
  
     
  
  
  
  
  
es image when the camera motion is uniform linear motion 
ly, (see Figure 3). These traces are straight lines if the optical 
oll axis of a camera is perpendicular to the camera path. The 
ne distance from the camera path to the target buildings is 
of able to be measured the slants of these lines. 
la It is necessary to prepare many cameras for wide area 
observation in an EPI analysis. In an EEPI analysis, all 
flows of stationary objects appear from FOE (Focus of Z 
Expansion) are picked up by the camera whose the optical 
axis is parallel to the camera path (see Figure 4). Thus a 
wide region can be observed at once by a small number of 
cameras. The obtained EEPI (see Figure 5) can be 
nd translated to EPI with simple geometric formulas. 
nd 
nd 
Is : : 
R Figure 3: Trace of feature point on EPI 
to 
gs 
lel 2.2 Construct of data 
a j 
ok The texture data for expression of urban scene are y imagedplane fort: 
at obtained from the image sequence. The slit spatiotemporal ; J 
lit plane image utilized as texture data integrate vertical lines view-point ~~ / 
or of both sides of each frame of image sequence toward the 
ed temporal axis (see Figure 6). 
of Let this slit image corresponds to distance data obtained 
as the EEPI analysis above, we call this depth map of urban ; 
scene. "Xx 
image-plane for EEPI 
Z(moving direction 
of viewer) 
Figure 4: Flow of stationary object under straight moving 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 913