Science Service System

Summary of Proposal MTH0135

TitleInterferometric Techniques for TerraSAR-X
Investigator Loffeld, Otmar - University of Siegen, Center for Sensorsystems (ZESS)
Team Members
Dr.-Ing. Brenner, Andreas - FGAN (Forschungsgesellschaft für Angewandte Naturwissenschaften e.V.), FHR (Forschungsinstitut für Hochfrequenzphysik und Radartechnik)
Prof. Dr.-Ing. Ender, Joachim - FGAN (Forschungsgesellschaft für Angewandte Naturwissenschaften e.V.), FHR (Forschungsinstitut für Hochfrequenzphysik und Radartechnik)
Prof. Dr. Loffeld, Otmar - University of Siegen, Center for Sensorsystems (ZESS)
Dipl.-Ing. Nies, Holger - University of Siegen, Center for Sensorsystems (ZESS)
Medrano Ortiz, Amaya - University of Siegen, Center for Sensorsystems (ZESS)
Dipl.-Ing. Gebhardt, Ulrich - University of Siegen, Center for Sensorsystems (ZESS)
M.Sc. Natroshvili, Koba - University of Siegen, International Postgraduate Programme Multi Sensorics (IPP)
Dr. Knedlik, Stefan - University of Siegen, Center for Sensorsystems (ZESS)
Dipl.-Ing. Walterscheid, Ingo - FGAN (Forschungsgesellschaft für Angewandte Naturwissenschaften e.V.), FHR (Forschungsinstitut für Hochfrequenzphysik und Radartechnik)
SummaryInterferometric SAR processing consists of individual processing steps applied in a sequence of well defined operations. Single look complex images must be co-registered, an interferogram must be formed (including eventually necessary prefiltering operations), the interferometric phase image must be unwrapped, taking into account the noisiness of the data, and, finally, the phase to height conversion requires very precise parameter estimates in terms of the interferometric baseline and position of the carrier platform. The works of this proposal aim at improving three individual key processing steps:
  • Orbit and Baseline Estimation
  • First we want to improve the given position and velocity data by a Kalman filtering and smoothing approach, which uses acceleration information at the actual satellite’s position for correcting and interpolating the orbit data. Besides the available orbit state vector measurements, the Kalman filter incorporates realistic models for gravitation and air drag, thus enabling very accurate orbit estimation, propagation and interpolation. The geometric baseline vector can then be calculated as the difference vector of two refined orbit positions, it can be further transformed into the interferometric baseline and then be further improved by using co-registration vectors, determined during the co-registration process.
  • Image Co-registration
  • Additionally to the classic correlation based methods, we will employ, compare and combine theoretical methods, based on information theory, such as mutual information, alignability in conjunction with automated bin size determination to achieve optimally co-registered images.
  • Phase Unwrapping
  • Our phase unwrapping approach is based on the optimal phase demodulation capabilities of nonlinear Kalman filters. Working on complex data, this approach additionally incorporates local phase slope information which is fused in an optimal way. The results will be compared with other state of the art approaches.
All the individual contributions will be developed in a modular way and implemented using "IDL". Hence they can be easily integrated into any interferometric processing chain.

Back to list of proposals

© DLR 2004-2016