Tal Abramovich 
  
  
Figure 4: GIS data (blue) and hypothesized road features (red and green) overlaid on an orthophoto. 
Marked GIS segment is matched with red image segments. 
0.6} 
04+ 
  
  
  
Figure 5: Resulted similarity values for two road segments. 
5 SUMMARY 
In this paper, a method for matching hypothesized road segments, extracted from an aerial image, and road 
entities in GIS, has been presented. The main idea is to quantify the similarity between these features. The 
matching is based on formulating the problem as a set of fuzzy rules. Using this methodology, a certain input 
does not have to “have” or “do not have” a certain property, but could “partially have” that property. This type 
of problem formulation is suitable for information extracted automatically from an aerial image. 
Preliminary results show that it is possible to obtain a similarity value between a hypothesized road segment in 
the image, and a road entity from a GIS. A comparison of the values obtained for a true match and false matches 
shows a difference, however there are a few matches with high values. 
  
22 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 
One 
Futu 
men 
aboi 
exar 
ACI 
This 
REI 
Ago 
geos 
Baur 
GIS 
Phot 
Heip 
Arch 
Zade 
IEEI