SAR 3-D imaging algorithm via Threshold Gradient Pursuit

Abstract

In recent years, the theory of compressed sensing has attracted great attention in radar field. As a typical iterative greedy algorithm, the Orthogonal Matching Pursuit (OMP) algorithm has applied to 3-D synthetic aperture radar (3-D SAR) imaging. But for large scene imaging, the OMP algorithm requires a huge computational time and space storage. The Gradient Pursuit (GP) algorithm has more advantages than the OMP algorithm in computation and space storage. However, both of the OMP algorithm and the GP algorithm need to set the sparsity level of the scene in advance, but the sparsity level in 3-D SAR imaging usually unknown. Aiming at the problem, this paper presents a 3-D SAR imaging approach based on Threshold Gradient Pursuit (TGP) algorithm. The algorithm uses the maximum minimum scattering coefficient ratio and the change rate of the scattering coefficient as the criterion for iterative termination instead of the sparsity level. Simulation and experiment results show that the proposed method cost less computational time compared to the OMP algorithm and has better performance than the GP algorithm at the same conditions.