On the applicability of compressive sensing on FMCW synthetic aperture radar data for sparse scene recovery

Compressive Sensing (CS) has proven its effectiveness for many applications by reducing the sampling rate and decrementing the acquistion time. This study evaluates if, how and under which conditions, CS can be applied on Frequency Modulated ContinuousWave (FMCW) Synthetic Aperture Radar (SAR) data for scenes containing only a limited number of scatterers. Two different approaches for the reconstruction are proposed and are evaluated on real data: (1) the CS reconstruction of the subsampled FMCW signal itself and (2) the CS reconstruction of the scattering coefficients of the scene. The limits of performance of CS are shown through simulated SAR data. The influence of the sparsity and the choice of a sampling scheme are evaluated in function of the number of samples needed for the exact reconstruction of the scene.