GEOMETRIC CALIBRATION OF THE STEREOSCOPIC CCD-LINESCANNER MOMS-2P
W. Kornus!, M. Lehner !, F. Blechinger?, E. Putz?
! German Aerospace Research Establishment (DLR), Institute of Optoelectronics, Optical Remote Sensing Division
D-82230 Wessling, Germany; e-mail: wkQoe.op.dlr.de, manfredQoe.op.dlr.de
? Daimler-Benz Aerospace AG (Dasa), Dornier Satellitensysteme
D-81663 Munich, Germany; e-mail: 100347.3010@compuserve.com
? Technical University Munich, Chair for Photogrammetry and Remote Sensing
D-80290 Munich, Germany; e-mail: putzOphoto.verm.tu-muenchen.de
Commision |, Working Group 3
KEY WORDS: Calibration, Scanner, Accuracy, Matching, Simulation, Geometric Scanner Calibration, Multi Scale Matching,
MOMS-2P/PRIRODA Mission
ABSTRACT
This paper describes the geometric calibration of the Modular Optoelectronic Multispectral Scanner MOMS-2P, which will
collect threefold stereoscopic panchromatic and multispectral imagery of the earth with 18 m and 6 m ground pixel size from
the PRIRODA module of the Russian space station MIR. The goal of the geometric calibration is the knowledge of the exact
object side image angles for a precise photogrammetric three-dimensional reconstruction of the earth's surface, considering all
geometric influences of the scanner like irregularities of the CCD-arrays, bending and relaxation of the focal plates, distortion
of the optics, etc. Prior to the mission the optical system has been calibrated in the laboratory of Daimler-Benz Aerospace
AG (Dasa), which is also the developer and manufacturer of the scanner. During the mission a geometric inflight calibration
will be performed in order to verify the calibration parameters on a routine basis. The first part of the paper describes the
lab-calibration, the equipment, the process and shows also the resulting calibration parameters. The second part deals with the
inflight calibration. The principle, the test areas and the mission scenario are outlined. An approach for multi-scale matching
using line scanner imagery and existing digital orthoimages is described. Finally the results of a simulation study are presented,
showing to which extend the camera parameters can be determined by photogrammetric means under realistic conditions.
1 INTRODUCTION locations are defined by two image angles around two orthog-
onal axes with respect to the optical axis of a camera or a
During the 10 day lasting D2 mission in spring 1993 the first scanner system. By means of a camera model, these angles
threefold along track stereoscopic imagery were recorded from normally are expressed in terms of image coordinates. For
space by MOMS-02 [11]. A series of photogrammetric eval- the precise transformation of image coordinates into spatial
uations were published, demonstrating its capability of gen- image angles, the camera model must accurately represent
erating photogrammetric products like digital terrain models the actual camera geometry implying its perfect knowledge.
(DTM) with an accuracy of 10 meters and less (e.g. [1], Since image coordinates are derived by digital measurement
[5]. [6]. [8], [10]). After D2, MOMS-02 has been refur- techniques with up to 0.1 pixel precision, the accuracy of the
bished and adapted to the PRIRODA environment for a re- geometric camera calibration must not exceed this value, if
fly on-board the Russian space station MIR. Besides the now the overall accuracy potential shall not be restricted by the
called MOMS-2P, the PRIRODA (Russian: nature) module system itself. In order to meet this requirement the process
contains a series of remote sensing sensors (see Figure 1) of the geometric camera calibration has been split into two
and will be docked on the MlR-station by mid April 1996. parts: a laboratory (lab-)calibration prior to the mission and
MOMS-2P consists of 5 lenses for simultaneous multispec- an inflight calibration accompanying the whole mission. The
tral and panchromatic 3-fold stereoscopic data acquisition inflight calibration is of highest importance, since the results
with a ground resolution of 18 m and 6 m from approximately of the lab-calibration proved to be reproducible in some cases
400 km height (see Figure 2). The Stereo module consists of only up to 0.5 pixel accuracy. Furthermore, temporal changes
3 lenses with one CCD-array each, providing a forward, a high of the camera geometry appeared in the MOMS-02/D2 evalu-
resolution nadir and an aft looking view. Additional 2 lenses ations indicating, that the lab-measurements may not exactly
with 2 CCD-arrays each enable multispectral imaging in 4 match the actual camera geometry in space.
spectral channels. The MOMS-2P/PRIRODA mission will
last at least 18 months and can be considered as preparation The main goal of the lab-calibration is the exact determi-
for a future operational and commercial MOMS-03 mission, nation of geometric irregularities of the single CCD-arrays,
planned for the end of this decade. The operational deriva- which hardly can be determined by inflight methods. The
tion of DTM and digital orthoimages (DOI) of wide parts inflight calibration is for the verification or change detection
of the solid earth's surface is a major goal of both missions. of the camera parameters and to determine the orientation
The photogrammetric evaluation is based on a measure- of the MOMS-2P camera axes with respect to the MIR co-
ment technique using intersecting spatial image rays, whose ordinate system in order to improve the knowledge about the
90
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B1. Vienna 1996
TV Came
IR Spectre
JSTOK- 1
VIS/NIR ¢
MOS-OI
MW Scan:
R-40
A
SAR Ar
Pri
absolu
also ti
This |
conclu
calibrz
sulting
the in
missio
identi
via an
and e
simule
camer
means
Exten:
in the
ule. T
