International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 Interna 
3. ENHANCEMENT FEATURES ON IMAGES 
give us different pieces of in formation about the same objects 
-— 
on the scene. 
  
Figure 4. Ground thrut data 
Areas under vegetation, bare land, stones and rocks are 
blending on the scene. The only artifacts on the scene are 
stone walls. In order to define the number of classes on the 
scene more easily, the existing images must be differentiate 
as well as possible. 
  
Figure 1. Part of DOF 
  
32. Pri 
Figure 5. Ground thrut data From th 
principa 
informa 
compon 
of the sc 
  
  
  
Figure 6. Ground thrut data 
3.1. Enhancement linear features and vegetation on scene 
For enhancement of linear features on R and G channel of 
VNIR mosaic filters Sobel are used. Locally adaptive contrast 
enhancement was done with Volterra/Unsharp filter. 
Vegetation on the scene was treated with Gram-Schmid filter 
on IR and R components of VNIR channesl, and filter TVI 
executed on the IR and R channel of VNIR mosaic. After Figure 
those enhancement there were 11 different images of the | 
same terrain produced (TIR, 3 VNIR components, 2 | 
Volterra/Unsharp filter on IR and R channels, 2 Sobel filter 
on R and G channels, 2 Gram-Schmid filter on 3 VNIR | 4.1. Sam 
components and TVI filter on IR and R channels). 
  
From th 
combinat 
without ’ 
recogniti 
created w 
and the s 
  
  
Figure 3. Geocoding TIR mosaic over VNIR mosaic 
1256