bul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004 
  
Result can be gotten from the experiment: 
double 1. When the measured error is 0, the targets" positioning 
r which error calculated is computation error. In flat area, 
horizontal precision is relatively high, but the vertical error 
targets’ is greater. In mountain area, the targets’ error mainly 
varying caused by the terrain relief. In this experiment, the 
horizontal error is smaller than 15 meters in X direction 
ie. the and in Y direction is smaller than 14 meters. And the 
; 10%, altitude error is very great. So this method is good for 
; 20%, horizontal positioning but not suitable for the altitude 
%, and poisoning. 
% The 2. Result can be found from the second and third group 
im, and of experiment (overlap is 20% or 30%) that the horizontal 
eriment error is basically unanimous, and mainly influenced by the 
terrain relief. The horizontal precision is great. Error both 
in X and Y direction are not more than 30 meters (in the 
limit of double maximum square error caused by the 
measured error which is 0.02 mm). When the overlap 
measure is as big as 60%, the horizontal precision reaches 
in cm grade, and the vertical error is about 1 meters. 
3. The targets' positioning error is up to the linear 
increase of measured error. 
4. CONCLUSION 
The following conclusion can be summarized through above 
experiment and analysis. 
I. When satellites’ position can be determined by GPS 
accurately, the targets’ positioning can be implemented. 
For the characteristic of the area adjacent images overlay 
can be used to correct the attitude of the images. The 
method stated in this paper is correct and feasible, and it 
offers a scientific approach of targets’ positioning with no 
GCP. 
2. In this approach, the precision of the photograph 
measurement is the main factor influencing the targets’ 
positioning. So the responding points can be found out 
with automatic match technology, in order to dispel the 
influence of the measured error as much as possible. 
3. The experiment of this method is still on the basis of 
the simulated data at present. So the next job is to utilize 
the true satellite remote sensing data and the aerial 
triangulation principle deepening the research on this 
model in order to form the practical method of the image 
orientation with few or even no GCP. 
  
  
  
20% 
113