Sparse aperture ISAR imaging via nonconvex low-rank and sparsity regularization

In inverse synthetic aperture radar (ISAR) imaging, traditional range-Doppler (RD) algorithm fails to get satisfactory imaging results in sparse aperture scenarios. Compressed sensing (CS) imaging methods are commonly used for sparse aperture radar imaging. However, only sparsity characteristic is used in CS-based methods. To improve imaging performance, other characteristics should be used. Range profile matrix exhibits low-rank characteristics as range profiles are similar during short ISAR imaging time, which can be employed in imaging. Therefore, a new joint sparse and low-rank imaging algorithm is proposed. To achieve superior imaging results, log-sum based nonconvex regularization rather than convex regularization is employed. Notably, low-rank and sparse problems based on log-sum regularization have closed-form solutions. Aternating direction method of multipliers (ADMM) algorithm is employed to solve the joint problem efficiently. Experimental results on measured data of Yak42 aircraft demonstrate that, compared with traditional methods, the proposed method achieves better imaging results in much less time. The imaging time is reduced by one to two orders of magnitude compared with L1/2+MC and SL0_LRLMS methods, and is less than half of that required by the convex regularization method. Therefore, the proposed method is an efficient sparse aperture ISAR imaging method with high imaging performance.