Compressive Sensing Analysis of Synthetic Aperture Radar Raw Data

This work addresses the use of compressive sensing to compress real Synthetic Aperture Radar (SAR) raw data. Due to the low computational resources of the acquisition platforms and the steadily increasing resolution of SAR systems, huge amounts of data are collected and stored, which cannot generally be processed on board and must be transmitted to the ground to be processed and archived. Although compressive sensing (CS) has been proposed and studied by a lot of researchers, almost none of them touches the real application of it. While, in this paper, we test the sparsity of the real SAR raw data (obtained by University of Kansas in Greenland, 2010), compress it using compressive sensing, and then recover the original signal using several CS recovery algorithms (Basis Pursuit, Matching Pursuit and Orthogonal Matching Pursuit), and compare these methods' performance. Simulation results are presented to prove the successful application of CS to real SAR raw data. When proper sparsity matrix is chosen, the real SAR data could be transformed to sparse signal. Using our designed algorithm, the positions and the exact values of the SAR raw data can be almost perfectly recovered with a very low MSE at a compression ratio of 1/8. This is of great significance to help us perform further research in the applications of CS to real SAR raw data.