International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Figure 7: Plot of the generated tie points in orbit 18
On the one hand there is the point determination of HRSC points
with three ray points and on the other hand there is the MOLA
DTM observation. Under these circumstances the differences be-
tween HRSC object points and MOLA DTM can be found. Fi-
nally, the MOLA DTM in these areas can be eliminated with a
robust adjustment.
orbit Xo[m] | Yo[m] | Zo[m]
bias 90.4 -64.6 -38.2
18 c 7.3 11.0 1.6
bias / a 12 6 24
bias -300.5 | -183.9 | -81.8
22 c 24.5 39.5 3.5
bias/ c 12 4.7 23
bias -12.1 -112.3 -41.2
68 G 10.7 16.7 6.7
bias/ c 1.1 7 6
Table 3: Accuracies of orbit determination (position bias)
orbit p[mgon] | w[Mgon] | x[mgon]
bias -51.1 -64.4 -6.2
18 a 0.3 lS 0.1
bias/c 194 4.2 63
bias -70.1 -26.3 -8.9
22 c 0.5 2.7 0.7
bias / o 164 10 13
bias -24.9 -12.1 -35.9
68 a 0.4 1.9 0.6
bias / co 64 11 57
Table 4: Accuracies of orbit determination (attitude bias)
Table 3 and 4 show the improved bias and bias accuracies o of
the positions and attitude for three orbits. In most cases the values
can be determined highly significant, because the bias accuracies
are lower than the bias values. In orbit 18 the planimetric accu-
racy o of the bias is better, because the terrain is rougher.
The accuracies of the coordinates of the object points for the or-
bits 18, 22, and 68 are shown in Table 5. They depend mainly on
two factors. First, there are the accuracies of the ray intersection
(Tab. 2) determining the accuracies within the orbit itself. Sec-
ond, there are the accuracies of the absolute orientation between
orbit and MOLA DTM (Tab. 3, Tab. 4). Thus, the precision of the
point determination is a combination of these both accuracies.
The absolute accuracies of the object points in all three dimen-
sions are less than 20 m for orbit 18 and 68. As a consequence of
the higher altitude the result for orbit 22 is slightly worse.
orbit | cx [m] | ev [m] | ez [m]
18 9.1 10.6 17.0
22 25.6 35.4 20.9
68 14.4 16.7 17.5
Table 5: Accuracies of HRSC points fitted to MOLA DTM
Finally, Table 6 shows RMS differences in Z between object co-
ordinates of HRSC tie points and MOLA DTM. In one case the
856
result is computed without DTM as control information and in
the other case with DTM information. The RMS differences in Z
between DTM and HRSC object points are in the range of 200 m,
After the bundle adjustment using DTM control information the
RMS differences in Z decrease by a factor of three. This means,
the adaptation of HRSC data to the MOLA reference system has
succeeded.
orbit 18 22 68
without DTM 177 m | 268m | 200m
with DTM 84 m 56 m 63m
Table 6: RMS differences in Z
5 CONCLUSION
The results show the potential of the image matching and bundle
adjustment approaches to achieve an improved exterior orienta-
tion with MOLA DTM as control information. The improvement
of the accuracy is significant. The position of the exterior orien-
tation increases from an a priori value of 1000 m to less than 20m
in all three dimensions (Tab. 5). The accuracy of the attitude in-
creases from 28 to 1-2 mgon in all angles. With the MOLA DTM
as control information the position and attitude can be improved
by an average factor of 30 - 50. Thus, after the bundle adjustment
the object coordinates of the tie points have a very high accu-
racy. Finally, there is a high consistency between HRSC points
and MOLA DTM, which constitutes the valid reference system
on Mars.
REFERENCES
Astrium, 2001. Mars Express — System Requirements Specifica-
tion. Vol. MEX.MMT.TN.0519, Astrium.
Carsenty, U., Flohrer, J., Jobs, D., Matz, K., Sebastian, I. and We-
semann, K., 1997. HRSC FMI and FM2 Calibration Document
(Final Report). Technical report, DLR, Institut für Planetenerkun-
dung.
Ebner, H. and Ohlhof, T., 1994. Utilization of Ground Control
Points for Image Orientation without Point Identification in Im-
age Space. In: International Archives of Photogrammetry and
Remote Sensing, Vol. 30, Part 3/1, pp. 206-211.
Ebner, H., Kornus, W. and Ohlhof, T., 1994. A Simulation Study
on Point Determination for the MOMS-02/D2 Space Project
using an Extended Functional Model. In: Geo-Informations-
Systeme, pp. 11-16.
Hechler, M. and Yañez, A. 2000. Mars Express — Consolidated
Report on Mission Analysis Issue 2.0. ESA.
Heipke, C., Schmidt, R., Oberst, J., Neukum, G. and the HRSC
Co-Investigator Team, 2004. Performance of Automatic Tie Point
Extraktion Using HRSC Imagery of Mars Express Mission. In:
International Archives of Photogrammetry and Remote Sensing,
Vol. 35.
Int
