stanbul 2004
|, coordinates
leviations of
d 0.09 pixel
X
- of 8555000
(D THREE
om the Otto-
m with the
rejection of
hreshold for
etry of high
three images
ism for the
third stereo
process was
ch results in
ightly better
tained from
d maximum
s more than
rence DEM
rom two
HMA)
ersection
three rays
662143
652764
9.65 94)
662076
-100 94)
5:16
9.89
-61.7
105.3
idistant grid
x 15m pixel
sses (forest,
ows that the
are best for
cital terrain
POT DEM.
rd deviation
ction.
M derived
52, HMA)
e different
ts)
Cities
Min. height difference [m] | -79.8 -81.0 -75.8
Max. height difference [m] 79.0 90.9 59.6
Intersection of three rays
Forest Open Cities
Areas
Mean height difference | 11.57 9.82 10.74
[m]
Standard deviation [m] 7.16 4.22 4.47
Min. height difference [m] | -122.0 -75.0 -106.1
Max. height difference [m] | 66.4 52.8 61.2
As shown in the accuracy analysis of the orthoimages the
channel HMA differs about 24 m in one direction with respect
to ground control points and keeping in mind the more complex
shift patterns in figure 2. This can be the reason for the only
slightly better results in the case of using three images for the
DEM production.
8. CONCLUSION
It could be shown that a stereoscopic evaluation of SPOT-HRS
data, only using ancillary data delivered by the image provider,
leads to an absolute accuracy of terrain heights in the order of
about 9 meter (mean height error), with standard deviations of
about 3 meter for high quality single points (two ray stereo
data) and 5 to 6 meter standard deviation for mass points of the
two and three ray stereo data in comparison to the reference
DEM. The relative and absolute accuracy for overall
comparison of the interpolated DEM (surface model) with the
reference DEM (terrain model) of course depends on land use
classes and terrain steepness, since image matching algorithm
depend on these features. For open areas a mean height error of
about 10 meter and a standard deviation of about 6 meter is
achieved, whereas a slightly better result (0.5 meter reduced
standard deviation) is obtained for three ray stereo data. The
expected improvement of three ray stereo data evaluation,
which offers better control mechanism for the image matching
procedure, probably is compensated by the lower performance
of the meta data (interior orientation) of the HMA nadir looking
channel.
The absolute accuracy can be improved by using ground control
points to reach a mean height difference of about 1 meter
(Reinartz 2004). The standard deviations can probably be
improved by DEM filtering, although not very large
improvements are expected.
Orthoimages can be derived to an absolute location accuracy of
| to 2 pixels (10 to 20 meter) without ground control, which is
in line with the performance specification. The shifts between
the orthoimages of the HRSI/2 stereo channels are highly
constant throughout the images, offering corrections with few
ground control points.
9. REFERENCES
S. Airault, P. Gigord, L. Cunin, E. Breton, A. Bouillon, R.
Gachet, M. Bernard, "Reference3d Location Performance
Review and Prospects", Proceedings of the ISPRS Workshop
High Resolution Mapping from Space, Hanover, Oct. 6.-8.
2003, IPI, Universität Hanover, on CD-ROM
A. Bouillon, E. Breton, F. De Lussy, R, Gachet, “SPOTS
Geometric Image Quality”, IGARSS 2003, Toulouse, 21.-25.
July 2003, IEEE International Geoscience and Remote Sensing
Symposium
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
J.P. Gleyzes, A. Meygret, C. Fratter, C. Panem, S. Baillarin, C.
Valorge: "SPOTS System overview and image groud
segment”, IGARSS 2003, Toulouse, 21.-25. July 2003, IEEE
International Geoscience and Remote Sensing Symposium
C. Heipke, W. Kornus, A. Pfannenstein, “The evaluation of
MEOSS airborne 3-line scanner imagery — processing chain and
results", Photogrammetric Engineering and Remote Sensing,
Vol. 62, No. 3, pp. 293-299, 1996
P. Nonin, S. Piccard, "Performance analysis of DEM automatic
extraction from SPOTS sensors", IGARSS 2003, Toulouse, 21.-
25. July 2003, IEEE International Geoscience and Remote
Sensing Symposium
M. Lehner, R.S. Gill, *Semi-automatic derivation of digital
elevation models from stereoscopic 3-line scanner data",
IAPRS, Vol. 29, Part B4 Washington, USA pp. 68-75, 1992
W. Linder, “Geo-Informationssyteme”, Springer-Verlag Berlin,
Heidelberg, Germany, ISBN 3-540-65276-0, 1999
Ru. Müller, M. Lehner, Ra. Müller, P. Reinartz, M. Schroeder,
B. Vollmer, “A Program for Direct Georeferencing of Airborne
and Spaceborne Line Scanner Images”, ISPRS, Comission |
Mid-Term Symposium “Integrated Remote Sensing at the
Global, Regional and Local Scale”, Denver, CO USA, 10-15
November 2002
Ru. Miiller, G. Palubinskas, P. Reinartz, M. Schroeder, V.
Amann, R. Stitter, “From Airborne Digital Raw Data to Image
Maps“ PFG Photogrammetrie, Fernerkundung, Geoinformation,
Germany, 4, 4, (2003), S. 317-326
P. Reinartz, M. Lehner, Ru. Miiller, M. Schroeder, “Accuracy
for DEM and Orthoimages derived drom SPOT HRS Sterco
Data without using GCP”, Proceedings of the ISPRS ux
Congress, Istanbul 12-23 July 2004
SPOT IMAGE, “SPOT Satellite Geometry Handbook”, S-NT-
73-12-SI, Edition 1, Rev. 0, Toulouse, France 2002
