Near-Field 3D Sparse SAR Direct Imaging with Irregular Samples

Abstract

Sparse imaging is widely used in synthetic aperture radar (SAR) imaging. Compared with the traditional matched filtering (MF) methods, sparse SAR imaging can directly image the scattered points of a target and effectively reduce the sidelobes and clutter in irregular samples. However, in view of the large-scale computational complexity of sparse reconstruction with raw echo data, traditional sparse reconstruction algorithms often require huge computational expense. To solve the above problems, in this paper, we propose a 3D near-field sparse SAR direct imaging algorithm for irregular trajectories, adopting a piece of preliminary information in the SAR image to update the dictionary matrix dimension, using the Gaussian iterative method, and optimizing the signal-processing techniques, which can achieve 3D sparse reconstruction in a more direct and rapid manner. The proposed algorithm was validated through simulations and empirical study of irregular scanning scenarios and compared with traditional MF and sparse reconstruction methods, and was shown to significantly reduce the computation time and effectively preserve the complex information of the scenes to achieve high-resolution image reconstruction.