N DR 
le 
al 
id 
d 
A MODEL FOR THE ORIENTATION 
OF ALONG-TRACK SATELLITE STEREOPAIRS 
Francelina A. Neto 
Dept. of Photogrammetry and Surveying, University College London 
Gower Street, London WCI1E 6BT, England 
ISPRS Commission IV 
ABSTRACT 
An orbital model for the orientation of along-track satellite linear array stereopairs is under development at 
University College London (UCL). The model is directed towards orientating images to be collected by future 
satellites such as ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) and QMI 
(Optical Mapping Instrument), expected to be launched in the late 1990s. 
Results presented here were derived from simulated data since no satellite data is currently available. That used 
in the tests covers an area in south-east France, from which SPOT data was already available in the Department 
of Photogrammetry and Surveying (P&S). The results give indications as to the expected accuracy of this 
model. 
Key Words: accuracy, aerotriangulation, mapping, space imagery 
1. INTRODUCTION 
The advent of SPOT in the late 1980s brought a new perspective 
into the study of cartography from space (Gugan et al, 1988; 
Dowman et al, 1988). SPOT is equipped with a pushbroom 
imaging system, with a nominal pixel size of 10 meters when 
operating in panchromatic mode (CNES, 1991), and high 
geometric accuracy. Its unrestricted commercial availability and 
high resolution made possible the use of SPOT data on an 
operational basis. 
Several approaches to modelling SPOT imagery have been 
developed, and some of these models were described and 
compared during the OEEPE (Organisation Europeènne 
d’Études de Photogrammetrie Experimental) test of 1989 
(Dowman et al, 1991). Although SPOT is equipped with 
sensors measuring the attitude angular changes with time, which 
are provided with the image header file, this data was not used 
in the OEEPL tests. A comparison of the main findings by 
different centres, using the same 10 control points over the same 
area, is summarised in table 1. 
Gugan et al, 1988, and Dowman et al, 1988, agreed that SPOT 
imagery may be suitable for both the production/revision of 
maps at scales scales up to 1:50,000 and the revision of some 
1:25,000 scale maps. 3D (Three Dimensional) information is 
obtained using SPOT side-looking stereopairs (figure 1), 
though this type of imagery has some disadvantages. Firstly, 
the two image; are taken at different times, usually separated by 
days or even months. Moreover, since the images are taken 
from different sides, with different illumination, identification of 
the points is more difficult. These effects may be reduced using 
along-track stereo imagery. 
315 
  
  
  
  
  
  
  
  
Research rms of errors in (m) 
centre H Plan 3D 
Hannover 6.4 13.5 14.9 
IGN 4.7 8.7 9.8 
Milan 11.5 16.5 20.7 
Queensland 9.4 12:3 18.5 
CCM 6.7 211 22.2 
UCL 7.3 16.1 17.7 
  
  
  
  
  
  
Table 1 - Precision obtained from orientating a common SPOT 
scene using different models (after Dowman et al, 1991) 
right 
image 
     
  
  
  
© 
direction 
of flight 
Figure 1 - Across-track stereo viewing 
2. ALONG-TRACK STEREO SYSTEMS 
Along-track camera systems have at least two optical systems 
simultaneously scanning the Earth’s surface, each system 
having a different along-track angle. À stereoscopic image is 
obtained using backwards and forward-looking pairs of images 
as shown in figure 2. The cameras collect images within