A Sparse Imaging Strategy of Moving Target in WasSAR System

In wide-angle staring synthetic aperture radar (Was-SAR), moving targets exhibit complex motion, leading to severe azimuth-range coupling. Therefore, accurately refocusing moving targets remains a key challenge in SAR ground moving target indication (SAR-GMTI). This paper proposes a sparse imaging strategy of moving target in WasSAR system. First, multi-channel interferometry and spatial spectrum estimation method are applied to achieve moving target detection and radial velocity estimation. Then, low-order compensation for translational motion is performed by airborne GPS and radial velocity, and the residual translational motion error is further corrected using a method based on Legendre polynomial fitting. Next, the range-Doppler (RD) algorithm is employed to generate a coarse image, which serves as the input to the compressed sensing (CS) model. The alternating direction method of multipliers (ADMM) is then adopted to solve the sparse signal recovery problem. Unlike traditional ADMM, the proposed method incorporates a two-dimensional spatially variant error matrix to correct phase errors during the iterative optimization process. And then, the RD domain image is projected to the North-East-Down (NED) coordinate system using two-dimensional interpolation. Finally, a trained network is used to adaptively repair the structural deficiencies or low reconstruction accuracy caused by inappropriate CS parameter settings. Moreover, the proposed method can obtain high quality images of maneuvering targets in the NED coordinate system, facilitating subsequent detection and recognition of moving targets. The Ku-band three-channel WasSAR results demonstrate the effectiveness of the proposed algorithm.