hul 2004 
m would 
à 
  
)-system 
camera 
stem 
n 
À 
1eras for 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
  
97/1963” 
pass was 
eras are 
gle from 
igle. The 
in 
y 
EB 
ce 
  
historic 
shows the 
anoramic 
width of 
2.1.2 Panoramic Camera Specifications: Panoramic cameras 
view only a comparatively narrow angular field at any given 
instant through a narrow slit. Ground areas arc covered by 
cither rotating the camera lens or rotating prism in front of the 
lens. The scale variation, called panoramic distortion, is a result 
of the cylindrical shape of the focal plane and the nature of 
scanning. Also, scan positional distortion is introduced in 
panoramic imaging due to forward motion of the aircraft during 
the time a scan made. Compared to frame cameras, panoramic 
cameras cover a much larger ground area. With their narrower 
lens field of view, panoramic cameras can produce images with 
greater detail than frame images (Lillesand, T. M, Kiefer, R. W. 
2000). 
2.2 SPOT-4 and IRS ID 
  
  
  
Acquisition | Spatial | Swath 
Date Res. Width 
SPOT-4 PAN 1998 10 m 60 km 
IRS-1D PAN 2000 58m 70 km 
  
  
  
  
  
  
Table 2.Summary of SPOT and IRS-1D satellite characteristics 
2.3 Study Area 
Study Area 
IRS1-D (pan) 
  
  
Figure 4. Localization of the study area and coverage areas of 
Corona, SPOT-4 PAN and IRS-1D PAN images 
The study area is situated in the European peninsula of Istanbul, 
located in eastern of Durusu (Terkos) Lake, northwestern of 
Bosphorus between 41?20' and 41°10' North and 28%43' and 
29°07" East (Figure 4). 
2.4 Creating georefenced bases for change detection 
Rectification of an image requires accurate and homogenously 
distributed ground control points. IRS-1D images were rectified 
on 1/5000 scaled orthophoto mosaic. To perform the 
rectification process, affine transformation method was used 
and the image was resampled by using nearest neighbor 
method. IRSI-D satellite images has been rectified on 
orthophoto mosaic using 40 control points with an RMS error 
of 0.6789 m. Likewise, SPOT4-PAN satellite imagery has been 
rectified on georeferenced IRS1-D images using 87 control 
points with an RMS error of 2 m. Consequently, the 
contemporary georeferenced basis was created for the scanned 
Corona panoramic film. 
2.4.1 Georeferencing scanned Corona panoramic film: In 
this study, the changes of coastal lines were detected. 
Therefore, most of the control points were measured on costal 
line zone on both the reference and input images. Naturally, the 
homogeneous distribution of the control points on the image has 
been taken in to account (Figure 5). 
  
Figure 5. Distribution of GCP’s on Corona image 
The ratio between the altitude of Corona satellite (185200 m.) 
and the maximum exaggeration of elevation difference on the 
study area (100 m.) was calculated, it was seen that the relief 
displacements because of the elevation difference, can be 
ignored. Therefore, the study area was assumed to be flat and 
the polynomial rectification methods were realized. 
To obtain best rectification result, three different methods were 
applied. These are affine, projective and rubber sheeting 
rectification methods. 
2.4.1.1 Rectification using Affine Transformation Method: 
The amounts of the residual errors were between ~10-20 m. in 
the areas which the control points were placed closely. 
However, the errors were within ~ 130 m. where control points 
set widely apart. These errors were occurred because of the 
panoramic structure of the Corona. 
2.4.1.2 Rectification using Projective Transformation 
Method : The full frame film of the Corona was rectified by 
using projective transformation method as a second step. The 
camera calibration report of Corona did not exist. Therefore the 
camera and film information were downloaded from internet 
and the fiducial mark coordinates were calculated. 
Extent of the film : Width : 756.92 mm 
Height : 55.372 mm 
Focal distance :609.6000 mm 
The interior orientation was performed by using fiducial mark 
coordinates with 0.27 pixel (8.74micron) accuracy. After 
performing the interior orientation, the exterior orientation 
parameters were calculated by referencing the rectified IRS1-D 
image and measuring 30 control points with 2.8672m RMS 
error (mx=3.504, my=1.597). 
The parameters are: 
Xo : 516121.585 m 
Yo : 4638002.163 m 
Zo : 220809.223 m 
Omega : -12.45414 ° 
Phi : 4.21308 ° 
Kappa : 12.38176 ? 
As a result of check point measurement on the rectified image, 
spatial errors were occurred between 100-250 m. 
UU LIII REPRE 
b 
i