left image(87.11.29) : : 
right image(87.11.30) 
537 satellite orbit 
  
Fig. 3 stereo model of satellite images 
table 1. major characteristic of SPOT photogrammetric film 
  
major level 1B level 1AP 
  
characteristic|left |right|left |right 
  
  
  
  
  
x(mm) 150.0/182.8|173.0|210.0 
y (mm) 150.0/|150.0/|171.4|171.4 
Scale 1/400,000 |1/350,000 
  
focal length | 2072.0(mm)|2370.5(mm) 
  
  
  
  
  
1 pixel size | 0.0250(mm)|0.0285(mm) 
  
The total amount of control points used is 23 and 13 of these 
are GCPs and 10 are check points. The distribution of the GCPs 
is as in Fig. 4. 
380000 eA ac ty 
  
"=== RIGHI 
AAAAA CONTR 
©0000 CHECK] POINT 
Masa si GROSS| ERROR 
360000- 
  
340000 
320000 -+ 
M:LATITUDE) 
Y(T 
300000 — 
  
  
280000 
  
  
  
260000 fer neat car a ve -- — — — — oH ——  —.— .-- 
180000 200000 220000 240000 260000 280000 300000 
X(TM:LONGITUDE) 
  
  
  
  
  
Fig. 4 distribution of GCPs 
The GCPs coordinates were acquired by ground survey and from 
1:50,000 topographic maps. Image coordinates were acquired 
using the Zeiss P2 Planicomp for level 1B ànd 1AP. For digital 
imagery level 1A, LINE/PIXEL of GCPs were observed in subpixel 
units through image processing methods. Initial values for the 
position of satellite position were interpolated from the 
header data and exterior orientation and polynomial 
coefficients were approximated and satellite height was 
adjusted considering the earth curvature, Also, for the 
jnitial value for the orientation of satellite, image 
orientation(y) was used as xo, inclined angle of the sensor 
was used as $o, and wo was set to 00. 
3.2 Determination of optimal polynomial for exterior 
orientation for each preprocessing level 
The approximation of dynamic exterior orientation of SPOT 
imagery is an important factor influencing the reliability and 
accuracy of adjustment system and is also an essential process 
in the establishment of DEM’s and in the production of 
orthophotos. 
In this study, varied forms of equation (2) as shown in table 
2, were used to determine the optimal polynomial of exterior 
orientation for each proprocessing level(table 2). 
table 2. analysis case for optimum polynomial of E.O.P. 
(exterior orientation parameters) 
  
case|NO. of parameters remarks 
E.O. P. 
  
1st order:position 
1st order:attribute 
1 12 Xo ,Yo,Zo ,Xo,00,90, 
Kiy,Kzy,Ksy, Ky, Koy 
  
2nd order:position 
1st order:attribute 
2 15 |casei*Kioy?,Ki1Y?,K12y? 
  
1st order:position 
2nd order:attribute 
3 15 |casei*kK7y? , Kay? Koy? 
  
4 18 |casei4kz7y? Kay? ,Koy? ,K10y?, |2nd order:position 
k11y?,Ki2y? 2nd order:attribute 
  
  
  
  
  
  
The input data for GCPs were acquired through ground 
Surveying and for image coordinates were acquired using the 
analytical plotters(leve 1B and level 1AP) and using the GCP 
module of the ERDAS image processing package in subpixel 
units(level 1A). The result of adjustment for each case 
according to each preprocessing level is in table 3. 
It can be seen from the table that for level 1B, a 15 
parameter polynomial is the optimal polynomial, where first 
order for satellite position and second order function of 
scanned lines are applied. For level 1AP and 1A, 12 parameter 
polynomial with first order function of scanned lines is 
optimal. 
Also, to analyse the the 3-D positioning accuracy of each 
preprocessing level, bundle adjustment was carried out for 
each case with the optimal polynomial, (table 4, Fig. 5) 
374