ISAR Super-Resolution Imaging Approach Based on Low-Rank Sparse Joint Constraint Under Low SNR

A super-resolution imaging approach for inverse synthetic aperture radar (ISAR) systems under the condition of low signal-to-noise ratio (SNR) is proposed based on the theory of low-rank and sparse matrix decomposition. The approach unites low-rank constraints and sparse constraints, while taking noise constraints into account. The regularization terms corresponding to the above constraints are added to the objective function of the ISAR super-resolution imaging optimization model. The constructed objective function is then solved using the linear alternating direction multiplier method with an adaptive penalty function to realize the ISAR super-resolution imaging under low SNR. The performance of the proposed approach under low SNR is verified by simulation and real data processing. Our proposed approach achieves better super-resolution imaging quality than those of commonly used algorithms including the orthogonal matching pursuit (OMP) algorithm, the split Bregman iteration (SBI) algorithm, and an approach using low-rank and sparsity priors (LRSPs) under the same low SNR, which is verified by comprehensively comparing the target-to-background ratio (TBR), the peak SNR (PSNR), and the image visual effect. Furthermore, the range of motion compensations errors that the proposed LSJA-LS approach can adapt to is also presented.