The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
1251
mrt r
f;'i
<*«
-T—
mt
Z.Wÿ-J(
r——
... itAAL.
O’!
*
J' r
<r*
ma
®T4
-
MM
am
«•»
«
» a
...
— yrf,
«B
.-- r ——-.
*n
i r *
1 OB
-
Figure 6. DLGs collected by stereoscopic measurements from
SPOT-5 HRS and HRG image stereo pairs. (A) is a DLG of an
alluvial flat area. (B) is a DLG of a hilly terrain area. (C) is a
DLG of a glacial landform area.
3.3 DOM
DOM can be automatically produced from the oriented HRG
imagery and DEM extracted from the HRS stereo pairs. The
generated image from HRG imagery is panchromatic. For
generating colour DOM, it is necessary to fuse HRG and SPOT-
5 multispectral imagery. Because of lack of the blue component
in SPOT-5 multispectral imagery, the paper proposes a method
to generate colour DOM by fusing the HRG and SPOT-5
multispectral imagery. The basic idea of the approach is that the
blue component is simulated from other components in
multispectral imagery firstly and then adapts the standard 1HS
method to fusing HRG and multispectral imagery. The method
is different from others. Orthorectification and fusion are
performed at the same time, so the step of image registration is
not necessary for fusion. However, the SPOT-5 multispectral
imagery orientation must be done before the orthorectification
and fusion. In order to orientating multispectral imagery, an
orientated HRG image is used as a reference image and
correspondences in the two images are automatically extracted
by image matching technique. The ground coordinates of these
correspondences are acquired from the orientated HRG image
and a corresponding DEM. Then the multispectral image is
automatically orientated by regarding the correspondences as
GCPs.
Figure 7. Images orthorectificated and fused by HRG images
and multispectral images. (A) is a fusion image of a hillock
region. (B) is a fusion image of farmland area. (C) is a fusion
of a mountain region
4. CONCLUSION
SPOT-5 images are extensively made used of in the national
project of Surveying and Mapping of the West Regions in
China. Firstly, with few GCPs, SPOT-5 images have good
orientation accuracy for mapping topographic maps at scaled of
1:50,000. The kind of capability is helpful for reducing the
numbers of ground control points surveyed in extremely
adverse environmental conditions of the west regions in China.
By using the Multi-photo Geometrically Constrained Matching
technique, DEMs or DSMs can be automatically extracted from
SPOT-5 HRS and HRG images. Colour DOMs are also
automatically produced from HRG images and multispectral
images. Because the physical geographic and natural
environments of the west regions in China are execrable, it is
not fully competent for the whole mapping tasks of the west
regions that only using SPOT-5 images. However, SPOT-5
