Spotlight SAR sparse sampling and imaging method based on compressive sensing

Abstract

Spotlight synthetic aperture radar (SAR) emits a chirp signal and the echo bandwidth can be reduced through dechirp processing, where the A/D sampling rate decreases accordingly at the receiver. Compressive sensing allows the compressible signal to be reconstructed with a high probability using only a few samples by solving a linear program problem. This paper presents a novel signal sampling and imaging method for application to spotlight SAR based on compressive sensing. The signal is randomly sampled after dechirp processing to form a low-dimensional sample set, and the dechirp basis is imported to reconstruct the dechirp signal. Matching pursuit (MP) is used as a reconstruction algorithm. The reconstructed signal uses polar format algorithm (PFA) for imaging. Although our novel mechanism increases the system complexity to an extent, the data storage requirements can be compressed considerably. Several simulations verify the feasibility and accuracy of spotlight SAR signal processing via compressive sensing, and the method still obtains acceptable imaging results with 10% of the original echo data.