International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B1. Istanbul 2004 
  
Figure 14 shows the general view of the generated DTM. The 
black areas inside the image are the places that matching 
process was unsuccessful. 
The report of the general mass point quality, used for DTM 
creation, is: 
Excellent % (1-0.85): 77.2222 % 
Good % (0.85-0.70): 8.0818 % 
Fair % (0.70-0.5): 0.0000 % 
Isolated %: 0.0000 % 
Suspicious %: 14.6960 % 
The number in the parenthesis shows the cross correlation 
coefficient. 
For evaluating the DTM, the generated DTM subtracted from 
the provided DTM of the Area. The provided DTM as it was 
mentioned previously doesn’t cover the entire images and the 
coverage is about 60-75%. Therefore, the evaluation has been 
done on the common area. Because of the non stability in the 
coastal line, the arca for the comparison is selected as Figure 8. 
The difference image is shown in Figure 15. 
In some areas, there are some gross errors to the mismatching. 
It means that the software has accepted two points as a pair but 
in the reality they are not. It is different form the regions that 
the software declares them as not matched place. In Figure 15, 
some of them are indicated by the red color. 
The result of the difference is: 
-Median = -9.8m 
-Mean = -21.0m 
-Standard Deviation = 57.4m 
The OrthobasePro v8.7 has the ability of the evaluation of the 
generated DTM with a reference DTM. The result of this 
evaluation is: 
Total number of DEM Points Used for Checking Vertical 
Accuracy: 1249240 
Minimum, Maximum Error: -788.4983 m, 333.8307 m 
Mean Error: -21.2591 m 
Mean Absolute Error: 25.7627 m 
Root Mean Square Error: 63.8070 m 
4. CONCLUSION 
The region of study has a high elevation difference, urban 
areas, water, and low relief, which make it useful for the 
scientific purposes for automatic DTM extraction evaluation. 
The result of this study could be summarized as follows: 
1- Both softwares had problems in high mountain areas 
as the changes in relief were very high. 
2- In a low relief and low mountains, both softwares 
acted well. 
3- The minimum number of GCPs recommended is 6 
per image or 6 common GCPs. 
4- Tie points should be selected in both softwares, as 
they increase the accuracy of the orbital modeling. 
5- The modification of OrthobasePro v8.6 model for 
SPOT 5 was not so successful. The given model in 
OrthobasePro v8.7 for SPOT 5 works well. 
6- Both generated DTMs by PCI OrthoEngine v8.2.3 
and ERDAS Imagine — OrthobasePro v8.7 have 
approximately the same accuracy. However, it seems 
that PCI strategy in automatic image matching acts 
better. 
7- The DTM refinement and editing by user is a very 
important task after the job is done. The reliability of 
the automatic generated DTM is not high. 
8- As the images arc taken in a short time difference and 
the look angle for stereoscopy is very appropriate, the 
stereo imaging geometry of SPOT 5 data have a high 
capability for height information extraction. Manual 
DTM extraction will have a high reliability. 
REFERENCES: 
Baudoin, A., Schroeder, M., Valorge, C., Bernard, M., 
Rudowski, V., 2004, The HRS-SAP initiative: A scientific 
assessment of the High Resolution Stereoscopic instrument on 
board of SPOT 5 by ISPRS investigators; ISPRS Congress 
Istanbul 2004 
ERDAS® LLC, 2002, ERDAS Field Guide, Sixth Edition, 
ERDAS® LLC, Atlanta, Georgia. 
ERDAS® LLC, 2002, ERDAS Imagine Software Help, 
ERDAS® LLC, Atlanta, Georgia. 
PCI Geomatica, Geomatica Software Help, 2003, PCI, 
Richmond Hill, Ontario. 
PCI Geomatica Website, Available: 
http://Www.pcigeomatica.com. (accessed 7 May 2004) 
APPENDIX I 
In this appendix the coordinates of GCPs, CPs and Tie points 
are given. Table A-1 gives the image coordinates of the tie 
points, which are used in PCI. Table A-2 gives the image 
coordinates of the tie points, which are used in ERDAS. Table 
A-3 gives the image and ground coordinates of the ground 
control points and Check points, which are used in both 
softwares. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Image | x | Image | y | Image2x | Image2 y 
ID (pixel) (pixel) (pixel) (pixel) 
AT0001 8351.5 1298.5 8422.5 1221.1 
AT0002 1155.3 2003.5 1228.5 3081.5 
A T0003 6035.5 3136.5 6107.4 3092.2 
AT0004 1295.5 4938.5 1368.5 50135 
AT0005 3479.5 4931.5 3552.4 4944.2 
AT0006 6060.5 4991.5 6132.6 4942 
AT0007 1181.5 6862.5 1255.4 6932.5 
AT0008 3481.5 6906.5 3554.6 6913.1 
AT0009 6093.5 7031.5 6165.8 6975.3 
AT0010 10794.5 7078.5 10865.5 6963.5 
AT0011 1128.5 8943.5 12022 8994.2 
AT0012 3550.5 8881.5 3623.5 8878.5 
AT0013 1307.5 10794.5 1380.5 10817.5 
  
  
  
  
  
  
550 
Table A-1. The image coordinates of the tie points used in PCI 
Inte