ISAR Imaging Based on Probabilistic Pattern-Coupled Sparse Bayesian Learning

In this paper we consider the inverse synthetic aperture radar (ISAR) imaging. To obtain high resolution target images, a novel probabilistic pattern-coupled sparse Bayesian learning (P-PCSBL) algorithm is proposed, which uses a probabilistic coupled prior model to represent the block sparse structural characteristics of ISAR images. The P-PCSBL algorithm is derived by variational Bayesian inference technique, where a decay factor is introduced to make the reconstructed signal sparser, thereby enabling the P-PCSBL to have the ability of suppressing noise. Simulation experiments demonstrate that the P-PCSBL has robust block sparse recovery performance and can obtain high quality ISAR images under low signal-to-noise ratio.