some improvement compared to case A and case B. For the
best case, however, the accuracy is still about 7” for all three
angles, which nearly is also reached in case A and case B
under the same conditions. Summarizing it can be stated,
1 strip 2 strips
Version || deo | dos | dro || dpe | dws | dk
100/20 7.1: 11:3 |. 6.9 7.2 9:0 |. 6.8
100/5 7.0 7.4 6.9 7.1 72 |.68
1000/20 6.9 8.5 6.7 7.1 8.0.1 6.7
1000/5 6.9 7.2 6.7 7.0 71 6.7
Table 11: Case C: Standard deviations ["] for single strip and
two crossing strips
that the results of the simulation study are very promising
and show, that the accuracy requirements for the geometric
calibration can be met also by inflight methods, if control
information is available with the expected accuracy. If two,
or more, crossing strips are simultaneously evaluated, the re-
quirements on the accuracy of control information can even
be relaxed. Since sensor displacements of 0.1 pixel seem to be
detectable, changes of the camera geometry, caused e.g. by
thermal effects, can be determined and later be considered in
the photogrammetric processing of the MOMS-2P imagery.
4 CONCLUSION
The geometric calibration of the stereoscopic CCD linescan-
ner MOMS-2P is subdivided into a laboratory calibration and
an inflight calibration part. In the lab-calibration the posi-
tions of about 15 pixels per CCD-array were determined with
an accuracy better than 0.1 pixel. From the calibration data
also locations of the single CCD-arrays relative to a reference
coordinate system were computed. They were derived twice
from independent measurements. In some channels these two
values showed systematic differences up to 0.5 pixels, i.e. the
determination of the relative location is not reproducible with
the required 0.1 pixel accuracy. Thus it can be concluded,
that by lab-calibration the shape of the single CCD-arrays
could be determined with sufficient high accuracy, their rel-
ative locations, however, still need to be verified by inflight
calibration.
The principle of inflight calibration is based on photogram-
metric bundle adjustment using a high amount of very accu-
rate ground control information. It could be shown, that the
automated GCP extraction by multi-scale matching of line-
scanner imagery and digital orthoimages is feasible. A first
test using SPOT imagery and orthoimages in scale 1:5000
and 1:25000 demonstrated, that more than 1000 GCP with
an accuracy better than 5 m can automatically be provided
for the inflight calibration.
The simulation study showed, that the inflight calibration of
the camera parameters and the orientation angles is possible
and the accuracy requirements for the geometric calibration
can be met. With the expected 1000 GCP and 5 m abso-
lute accuracy of the camera position, the focal lengths of all
three lenses can be determined with up to 3 um, the sensor
displacement with up to 0.1 pixel, the sensor rotation with
up to 2.4" and the rotation angles with up to 7" accuracy.
If two, or more, crossing strips are simultaneously evaluated,
98
the requirements on the accuracy of control information can
even be relaxed.
In order to analyze the impact of temporal and thermal influ-
ences on the camera geometry, the inflight calibration is to
perform as often as possible during the commissioning phase
and should later be repeated in a reasonable time interval on
a routine basis.
REFERENCES
[1] E. Dorrer, W. Maier, V. Uffenkamp: "Analytical kine-
matic sensor orientation of MOMS-02 linear array stereo
imagery", Integrated Sensor Orientation, Colomina |,
Navarro J. (Eds.), pp. 261-273, Wichmann Verlag,
Karlsruhe, 1995.
[2] H. Ebner, W. Kornus, G. Strunz, O. Hofmann, F.
Müller: “A simulation study on point determination
using MOMS-02/D2 imagery”, Photogrammetric Engi-
neering & Remote Sensing, Vol.57, No. 10, pp. 1315-
1320, October 1991.
[3] H. Ebner, W. Kornus: “Point determination us-
ing MOMS-02/D2 imagery”, Conference Proceeding
IGARSS, Vol. Il, pp. 1743-1746, Helsinki, Finland,
1991.
[4] H. Ebner, W. Kornus, T. Ohlhof: "A simulation study on
point determination for the MOMS-02/D2 space project
using an extended functional model”, Int. Archives of
Photogrammetry and Remote Sensing, (29) B4, pp.
458-464, Washington D.C., USA, 1992.
[5] C. Fraser, D. Fritsch, P. Collier, J. Shao: “Ground point
determination using MOMS-02 earth observation im-
agery', Proceedings of the 37th Australian Surveyors
Congress, Perth, Australia, April 13-19, 1996.
[6] W. Kornus, H. Ebner, C. Heipke: "Photogrammet-
ric point determination using MOMS-02/D2 imagery",
Proceedings of SPIE Conference on Remote Sensing and
Reconstruction for 3-D Objects and Scenes, pp. 115-
125, San Diego, USA, 9./10. 7. 1995.
[7] M. Lehner: "Triple stereoscopic imagery simulation and
digital image correlation for MEOSS project" , Proceed-
ings of ISPRS Commission | Symposium, pp. 477-484,
Stuttgart, 1986.
[8] M. Lehner, W. Kornus: “The photogrammetric evalua-
tion of MOMS-02/D2 mode 3 data (Mexico, Ethiopia)",
Proceedings of SPIE Conference on Remote Sensing and
Reconstruction for 3-D Objects and Scenes, pp. 102-114
San Diego, USA, 9./10. 7. 1995.
[9] M. Lehner, W. Kornus: “Digital photogrammetric pro-
cessing of MOMS-02/D2 imagery”, ISPRS Congress -
Commission Ill / WG 2, accepted as poster, July 9-19,
1996.
[10] J. Schiewe, Y. Wang: "MOMS-02/D2 data for the
generation and revision of cartographical databases",
Proceedings of MOMS-Symposium, Cologne, Germany,
July 5-7, 1995.
[11] P. Seige: "MOMS: A contribution to high resolution
multispectral and stereoscopic earth observation from
space", Photogrammetric Week '93, Fritsch D., Hobbie
D. (Eds.), pp. 109-120, Wichmann Verlag, Karlsruhe,
1993.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B1. Vienna 1996
TR
KEY
ABST
Transp
aerial |
test the
High
photog
photog
photog
naviga
using |
can be
In pr
atmos]
mover
or pro
lower
quality
neede
The 1
resolu
or the
(Hakk
has be
most |
atmos
spatial
The o:
or im:
contra
in nu
imagi
qualit
at diff
an ot
photo
neede
In sui
experi
The i
they
target
exper
