The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
1248 
and the DORIS instrument, for greater orbital accuracy are 
contained in SPOT-5 System. 
SPOT-5 images have been taken as the main data source in the 
national project of Surveying and Mapping of the West Regions 
in China. Applications of SPOT-5 imagery in the project are 
introduced in the paper. 
2. IMAGES ORIENTATION 
For the orientation of the HRS and HRG images there are two 
approaches. The first one is a rigorous sensor model for 
pushbroom linear array sensors (Poli, 2003). The alternative 
approach is based on the RPC (Rational Polynomial 
Coefficients) model (Zhang et al., 2004). The principle of the 
latter is as the following. At first 3D grids of points were 
generated for a set of elevation levels from the image-space 
coordinates, using the given camera model parameters, the 
ephemeris and the attitude data attached in the metadata file. 
Then the RPC coefficients were determined for each image by a 
least-squares approach and without GCPs. After the RPC 
generation, a block adjustment was performed in order to 
estimate 6 parameters for each image (affine transformation) to 
remove remaining systematic errors. In the project of Surveying 
and Mapping of the West Regions in China applies the latter 
approaches for SPOT-5 images orientation. 
Two experiments were carried out to evaluate whether the 
SPOT-5 images are competent for cartographic maps at scaled 
of 1:50,000. The first experiment applied two SPOT-5 HRS 
stereo image pairs in two adjoining orbits to orientation for 
accuracy evaluation. One covers the ground extension about 
120><580 km 2 and another covers about 120x508 km 2 . A total of 
77 object points measured with surveying methods were applied 
for the bundle adjustment of the RPC model for SPOT-5 HRS 
images in the experiment. 
Figure 1 HRS stereo image pairs and GCPs applied for 
accuracy evaluation in first experiment 
In the experiment, different schemes of bundle adjustment were 
carried out by selecting different number of Ground Control 
Points (GCPs) and Check Points (CPs) from the total 77 object 
points, and compared the orientation accuracy in these schemes. 
The results are given in the following table 7. 
GCP 
s/CP 
GCP RMS 
Residuals (m) 
CP RMS 
Errors (m) 
s 
X 
Y 
Z 
X 
Y 
Z 
0/77 
22.35 
42.63 
7.14 
4/73 
4.07 
0.67 
1.42 
6.59 
6.41 
4.94 
6/71 
0.71 
1.20 
0.47 
6.34 
5.48 
4.01 
77/0 
5.65 
4.84 
3.64 
Table 1. Results from the least-square bundle adjustment of the 
RPC model for SPOT5 HRS stereo-pairs. 
The last experiment applied a SPOT-5 HRS stereo pair and a 
SPOT-5 HRG image covering the same ground area and carried 
on the least-square bundle adjustment of the RPC model for the 
HRS stereo pairs and the HRG image. 
Figure 2. HRG image and distribution of tie points and GCPs 
In the experiment, a total of 48 object points measured with 
surveying methods were applied for the different tests on the 
bundle adjustment of the HRS and HRG stereoscopic pairs. 4 of 
these object points distributed at the bound of corresponding 
area were selected as GCPs and others were selected as CPs. 
Two orientation schemes were carried out and compared the 
results of two schemes. In the first scheme only the SPOT-5 
HRS stereo images are applied for bundle adjustment. In 
another scheme the SPOT-5 HRS stereo images and HRG 
image were applied for bundle adjustment. 
GCPs 
CPs 
X 
Y 
Z 
X 
Y 
Z 
HRS 
5.177 
2.008 
2.151 
5.213 
3.439 
4.248 
HRS 
& 
HRG 
3.165 
2.401 
1.047 
2.764 
2.220 
2.467 
Table 2. Result from the least-square bundle adjustment of the 
RPC model for HRS and HRG. 
The results of experiments indicate that SPOT-5 images are 
competent for mapping topographic map at scaled of 1:50,000 
in West Regions of China. 
3. APPLICATIONS 
Once SPOT-5 images are orientated, coordinates of objects in 
object-space are able to be extracted from one or more stereo 
image pair. Digital Line Graph (DLG), Digital Elevation Model 
(DEM) or Digital Surface Model (DSM) and Digital 
Orthorectification Map (DOM) are classical products of digital 
photogrammetric processing. The generations of DEM/DSM, 
DLG and DOM from SPOT-5 images are introduced in the 
following sections. 
3.1 DEM/DSM 
Digital Surface Models (DSMs) are often referred to as the