In the second method for extraction 
three-dimensional coordinates out of the SPOT 
stereo pair, ground coordinates of check points 
are computed by the intersection of oriented 
rays after applying space resection. Computed 
heights are adjusted by using a second-order 
polynomial. The effects of observation errors, 
in measured coordinates, on computed heights are 
tested by applying another correction equation 
which takes into consideration these errors. 
Root mean squares of residuals at check points 
are computed, Table(3). 
Table (3). Root Mean Squares Values of Residuals at Reference 
Points Computed after Applying the Space Resection- 
Intersection Method. 
  
  
  
  
  
  
  
Number of Number of Number of Root Mean Squares (meter) 
Reference Orientation Control 
Points Elements Points Rx Ry Rz Rz' 
6 61 85 21 -- 
10 45 77 26 -- 
4 15 59 89 13 14 
20 52 81 10 20 
32 50 81 09 09 
32 
10 45 40 15 -— 
15 34 46 13 13 
6 20 35 43 12 10 
30 34 37 10 12 
32 33 25 09 09 
6 50 67 20 -- 
10 52 65 21 28 
4 15 43 60 20 22 
23 42 59 19 20 
23 
6 46 24 22 10.0 
10 45 23 18 2.3 
6 15 27 20 15 1.7 
20 27 19 14 1.7 
23 25 19 14 1.6 
* Rx Root Mean Square Error in Eastings, 
* Ry Root Mean Square Error in Northings, 
* Rz Root Mean Square Error in Heights, and 
* Rz' Root Mean Square Error in Heights after Reducing 
Effects of the Observation Errors. 
4- DISCUSSION AND CONCLUSIONS 
Obtained results indicated that modified 
relative orientation method, where changes with 
time in base components are represented by 
polynomials, reduces Y-parallax considerably. 
However, by applying three-dimensional 
polynomials for the transformation of model 
coordinates into ground coordinates, obtained 
accuracy are ,approximately, identical 
regardless of the used orientation method. 
288 
Obtained heights accuracy in the two methods 
of stereoscopic analysis are ,almost, equal and 
compatible with accuracy obtained in other 
investigations. For images with low to medium 
contrast heighting accuracy of 10 m can be 
obtained. With high contrast images and, 
accordingly, with less observational errors in 
image coordinates ,accuracy of heights will be 
limited mainly by the precision of ground 
coordinates and can be increased to about 2 m. 
So, it is important to include image enhancement 
in the photogrammetric processing of the SPOT 
images. 
The first procedure for the extraction of 
three-dimensional coordinates, the two steps 
orientation, is more simple to apply on 
analytical plotters and needs less computer time 
and storage. 
The second approach ,the one step orientation 
method, is flexible and allows for more 
parameters to be included in the adjustment. It 
requires more computer time and needs more 
storage capacity. 
  
REFERENCES 
DIEFALLAH, M.A.M. (1990) : 
Heighting Accuracy of SPOT Imagery. 
ISPRS-Commission (IV) Conference, Tsukuba, 
JApan, May-1990. 
DOWMAN, I.J. (1986) : 
Techniques for Mapping and Map Revision 
Using Digital Imagery from Space. 
International Archives of Photogrammetry and 
Remote Sensing. Vol. 26, pp. 529-537 . 
DUCHER, G. (1980) : 
Cartographic Possibilities of the SPOT and 
Spacelab Projects. Photogrammetric Record, 
10 (56), PP 167-180. 
HARTLEY, W.S. (1988) : 
Topographic Mapping with SPOT-1 Data. A 
Practical Approach by the Ordnance Survey. 
Ordnance Survey Report (1988). 
KONECNY, P. (1987) : 
Evaluation of SPOT Imagery on Analytical 
Photogrammetric Instruments. Photogra- 
mmetric Engineering and Remote Sensing, Vol. 
53, No.9, PP. 1223-1230. 
RODRIGUEZ, V. (1988) : 
Evaluation of Stereoscopic Accuracy of the 
SPOT Satellite. Photogrammetric Engineering 
and Remote Sensing, Vol. 54, No.2, PP. 217- 
221. 
THOMPSON, E. H. (1961) : 
Corrections to X-Parallaxes :Numerical 
Analysis. The Photogrammetric Record , 6 
(32), pp- 202-210. 
WELCH,R. (1983) : 
Imapact of Geometry on Height Measurements 
from MLD Digital Image Data. Photogra- 
mmetric Engineering and Remote Sensing, Vol. 
49, No. 10, pp. 1437-1441. 
(C) CY CI CP fru f) m Sets CCS m (n UT Cad "-— 
-—