In order to display the advantages of multi 
image matching, only the middle part of the 
testfield was processed. Figure 5-1 and 5-2 
show, that different areas are covered in the 
different images. 
The digital multi image orthophoto, on the other 
hand, reconstructs the complete field without 
any covering areas (Fig. 5-4). 
  
Figure 5-4 : A sector of the multi image 
orthophoto 
5.2 Example 2_: Natural object 
A natural object from the field of architecture 
photogrammetry was selected as a second example 
the tessellated floor in the cathedral of 
Siena (Italy). 
A Rollei 6006 camera was used to take the 
photographs, and a Rollei RS 1 scanner digitized 
the photographs /SINNREICH 1989/. The parameters 
of the interior and exterior orientation were 
defined in preparatory measurings. For this 
example, the size of the surface elements was 
1,5 x 1,5 mm?, the size of the mesh was 15 x 15. 
Figure 5-5 shows the different sections of the 
mosaic. This example proved that natural 
textures also supply sufficient information for 
the computation of orthophotos (see Fig. 5-6). 
    
  
  
  
  
   
  
    
feu 
5 
F 
    
  
igure 5-5 : Sections of the mosaic 
  
Figure 5-6 : Surface meshes of the single 
images and resulting mesh 
The overall result of this multi image matching 
is satisfactory (see Fig. 5-7). As the number of 
white sectors in the figure indicates, there are 
only very few failures left after the 
computation. However, more appropriate 
configurations of the recording stations and/or 
photographs of higher quality should diminish or 
completely eliminate the failures. 
  
Figure 5-7 : A sector of the multi image 
orthophoto