Abstract:

Under sparse aperture conditions, some problems arise with inverse synthetic aperture radar imaging such as low azimuth resolution and susceptibility to noise. To solve them, the two-dimensional sparseness of a target is used to transform the imaging problem into the sparse signal reconstruction problem under the multiple measurement vectors model. The zero norm-least mean square algorithm is processed in parallel to improve the efficiency. The optimal step-size formula is used instead of the fixed step-size to avoid the influence of the improper step-size on the convergence speed and accuracy. And the smoothed zero norm is introduced to approximate the zero norm to improve the reconstruction accuracy and noise immunity ability. In comparison with existing methods, the proposed algorithm can obtain a clearer target image, which is robust to noise and requires less computation. The effectiveness of the proposed method is verified by simulation and the real data processing result.

Key words: inverse synthetic aperture radar, compressed sensing, sparse aperture, adaptive filtering, smoothed zero norm, multiple measurement vectors