2D high-resolution ISAR imaging by joint using matrix completion and compressed sensing

In practical stepped frequency radar imaging application, due to the change of environment and radar multi-mode, the missing number and position of sub-pulses usually randomly distributed in each pulse group. Although compressed sensing (CS) method can handle this 2D recovering model by matrix vectorization, the huge computational complexity makes it difficult to implement on a normal personal computer. In this paper, a new 2D high-resolution ISAR imaging approach by jointly using matrix completion (MC) and compressed sensing (CS) is proposed. Firstly, the MC method is used to estimate the randomly missing elements based on the low-rank property of echo data matrix. Second, we further use CS method to handle the 2D imaging task based on the recovered full data matrix directly. At last, a high-resolution ISAR image can be obtained. Compared with competing methods, it has an improved imaging performance under low sampling rate conditions, and can effectively elevate the recovery performance of anti-noise.