Four-dimensional SAR imaging algorithm using Bayesian compressive sensing

Abstract

The compressive sensing (CS) based 4-D synthetic aperture radar (SAR) imaging method performs well in the case of high signal-to-noise ratios (SNR). However, in the presence of strong noises, the performance of CS-based method degrades and the number of false targets increases rapidly. In this paper, a novel 4-D SAR imaging method is proposed based on Bayesian compressive sensing (BCS). Assume that the target scattering field follows the Cauchy distribution, the 4-D SAR imaging is transformed into signal reconstruction via maximum a posteriori estimation. In addition, Poisson disk sampling is utilized to generate the radar positions of 4-D SAR in the baseline-time plane. Experimental results show that the proposed method is capable of effective suppression of the noise by exploiting the sparseness prior distribution of the image scene, and a well-focused image could also be achieved even under the condition of low SNR.