Interferometric Synthetic Aperture Radar Signal Processing

This thesis demonstrates the signal processing of interferometric synthetic aperture radar all-sided, especially in the challenging problems such as coregistration, denoising filtering and two-dimensional phase unwrapping. Based on the whole project of space borne INSAR signal processing, the airborne interferometry system is designed and experimented.First, the frequency shift between the two single look complex images is discussed and the effect of baseline and topography slop is analyzed. The difference of the incidence angle is uncovered as the essence reason of frequency shift. The author makes use of the range frequency spectra to estimate the frequency shift and designs non-symmetrical filters to increase the coherence of the signals.Since accurate coregistration is the interferometry element, several conventional registration methods are dissected and compared, which include correlation-based algorithm, average fluctuation algorithm and maximum spectra algorithm. According to the residue characteristic, a new registration method we named residues estimate algorithm is presented. The new algorithm can not only improve the registration accuracy, but also reduce the compute complexity.After investigating statistical and distributing property of the phase noise, the application of average filter and median filter in the phase denoise is canvassed and the defect of the above filters is pointed out. The paper focus on local statistics filter combined with the MUSIC algorithm and the melioration method is presented based on the LS filter.Phase unwrapping is the core technology that enables radar interferometry. The author digs into various unwrapping algorithms and gives a thoroughly compare through analyses and experiments. At the same time, a quality standard to measure unwrapped phase is established, based on which that a post-quality-validation projectintegrating multi-algorithms is presented. Experiment results confirm the feasibility and advantage of this method.Finally, the thesis describes the experimental study on repeat-pass airborne SAR interferometry. The key technologies includes airline control of aircraft, baseline determination, motion compensation, signal processing. A motion compensation method based on INS/GPS data has been proposed and implemented. An interpolation approach in range and azimuth dimensions is used to eliminate the effect of non-parallel baseline on correlation. Some resultant interferograms demonstrate the validity of the interferometric L-SAR. Suggestions of future experiments are put forward according to the results.