Error modelling for SAR interferometry and signal processing issues related to the use of an encoding SAR transponder

The central research themes addressed throughout this thesis are the following: • statistical modelling of atmospheric path length and phase unwrapping errors in repeat-pass SAR interferometry • derivation of a suitable focusing algorithm and performance in a random scene of a pulse-to-pulse encoding SAR transponder Given the state-of-the-art of interferometric SAR techniques, application fields for the former are error prediction for the generation of height and displacement maps with a minimum data set, data selection within multi-interferogram frameworks and interpolation of atmospheric error corrections obtained with external systems such as GPS and imaging spectrometers. In this thesis closed form models for the second order statistics (structure function) of tropospheric path length and phase unwrapping errors are derived. As an application, the derived models are used within a mathematical framework developed at Denmark’s Technical University (DTU), Electromagnetic department (EMI), to predict the error in several ERS-tandem Digital Elevation Models (DEMs). An SRTM DEM is used as a reference. The latter investigations in the above bullet-list address a specific type of artificial reflector. In the context of SAR interferometry, artificial reflectors may be exploited in tie-point baseline calibration methods as well as in Persistent Scatterer techniques. A specific transponder architecture is considered in this thesis, namely a pulse-to-pulse BPSK encoding one. A focusing algorithm is proposed and validated on an ERS data set containing prototypes built at DLR (German Space Agency) as well as on simulated data. The interaction of the encoded transponder signal with that due to backscattering from non-encoding objects is analysed. Equations enabling assessment of design trade-offs are derived.