Full text: Close-range imaging, long-range vision

  
  
  
    
            
DT EI T d 
19" century. 
— 1 I 
4 ode : 5 a T $e 
Bauakademie, originally constructed under Schinkel in the 
F igure 1: Images of the reconstructed corner of the destroyed 
  
e Apparent identity of repeated features on facades, like 
brick edges. 
e Large differences in the images due to different points 
of view. 
e Large differences in the images due to different image 
acquisition times. 
e Large differences in the images due to different 
  
  
backgrounds along edges of the discontinuous 
surfaces. 
| Feature based matching techniques may also be mentioned. But ee 
| also feature based matching approaches may fail due to one or Fig 
| more of the following reasons: 
e  Insufficiently sharp defined features 
e Apparent identity of repeated features det 
e Different distribution of features due different points par: 
of view rest 
| e Horizontal linear features are difficult to determine the 
| due to images acquired from the same height. and 
use 
othe 
2. IMAGE MATCHING part 
| bety 
To cope these challenges a few principles have to be used in | acq 
image matching techniques. One is to use a hierarchical 
matching approach. An other principle is to start from a 
building model consisting of plane faces and maximum positive 
and negative deviations from these planes. This data may be 
derived from a sketch or a preliminary CAD model. 
It is recommended to use at least pre-oriented images which 
allows an object space Least Squares Matching (LSM) (Helava 
1988). The orientation of architectural images is usually not a 
big problem and may be based on interactive measurement of a 
few control points and about 10-20 tie points per image. At | 
least two control or tie points per facade should allow the 
definition of the usually vertical projection planes. 
  
Figure 2: DSM derived from a CAD Model 
1.4 Problems of Image Matching Techniques 
Image matching seems to be a suitable solution. But the area 
based matching approach fails in a lot of situations due to one 
or more of the following reasons: 
2.1 Knowledge on the Surface Structure 
e Lack of texture on facades, especially on plastered : 
faces, The knowledge of the orientation parameters and approximate 
building boundaries allows the limitation of search areas, the 
  
—480— 
  
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.