The International Archives oj the Photogrammetry, Remote Sensing and Spatial Information Sciences. Voi. XXXVII. Part B7. Beijing 2008
Fig. (2): 3D view of glacier with DEM and Quickbird image color
composite (3,2,1)
4.3 Displacement computation using cross-correlation of
multi-temporal imagery
In this work ,a pixel-based image matching approach is used to
derive metric information on glacier displacement .Two well co
registered orthoimages are matched through correlation analysis.
The general principle of offset computation by correlation is the
following: around every reference image point, for which one
wants to compute the local offset with a secondary image, a so-
called rectangular “vignette” is constituted, centered on the
reference point. Having a rough knowledge of the local offset,
we can center the search on the secondary image around the
estimated position. We iteratively build a small vignette on the
secondary image, which has the same size as the primary one,
and which travels through an area called “research area” .The
size of this research area is related to the uncertainty on the
position in the secondary image(H.Vadon et al.,2001).
In order to measure correlation rates, we used weighted correlation
co-efficient which is calculated as follows :
X X M w, )(g 5 (*,, y, ) - g’ )(.g' (*,, yj ) - g' )
p* ■„ „ „ ——r
[Z J 'Z i ™(x l ,yj)(g 3 (x n y J )-g s ) 2 .'£'£ i w(x l ,y J )(g'(x l ,y J )-g') 2 ] 2
1=1 7=1 1=1 7=1
First, correlation rates are computed for every vignette position in
the research area. The best correlation corresponds to the position
in the research area of the secondary image that maximizes the
radiometric resemblance between the two vignettes. Then sub
pixel search with similar principle is iteratively performed until
the required localization accuracy is reached and afterwards offset
computation is done(H.Vadon et al.,2001).
After correlation analysis, the displacement map is produced.
5. RESULTS AND DISCUSSION
Fig. (3) Shows the horizontal displacement of the ground surface
of Alamchal glacier .On this glacier , no overall displacement
measurement had ever been performed.
This research shows a significant retreat in glacier front over a
long period of time, indicating a decrease in glacier reservoir.
Furthermore, the results demonstrated the occurrence of major
surface displacements in the central region of the glacier due to
both accumulation and ablation.
In the course of this study, some problems appeared. No
displacement could be measured for areas in shadows and at high
elevation on the accumulation zone, even with a low gain, the
noise remained noticeable.
4gyooo 4ggooo 4ggooo
Figure. (3): displacement image of glacier displacement
4 98 000
4 99
1
4gyooo
Figure. (4): 3D view of displacement of glacier
The apparent displacement of the shadows is negative and oriented
mainly parallel to image lines. The strength of the correlation is
weakened by a secondary correlation peak due to the shifted
shadows. However, on the glaciers, this shadowing effect does not
cause a systematic error in our measurements. Shadows are only
created by the surface debris and the surrounding mountains. The
best way to ensure a limited bias is to retain only the
measurements with the highest correlation coefficient. The effect
could be larger for other applications such as landslides, especially
if images are acquired during different seasons.