Efficient MMW Image Reconstruction Algorithm Based on ADMM Framework for Near-Field MIMO-SAR

Xu Chen,Chenggao Luo,Qi Yang,Liuxiao Yang,Hongqiang Wang

DOI: https://doi.org/10.1109/tmtt.2023.3297262

IF: 4.3

2023-01-01

IEEE Transactions on Microwave Theory and Techniques

Abstract:In recent years, various fast imaging algorithms have been developed for near-field millimeter-wave (MMW) sparse multiple-input–multiple-output synthetic aperture radar (MIMO-SAR) system, and most of them are implemented through fast Fourier transform (FFT). However, these algorithms have strict requirements on the uniformity and sparsity of MIMO arrays, and hardly guarantee high-quality image reconstruction under a large dynamic range. This article proposes an efficient MMW image reconstruction algorithm based on alternating direction method of multipliers (ADMM) framework. First, the forward sensing operator is constructed according to the existing FFT-based imaging model. Then we substitute it into the traditional ADMM framework, and at the same time update the iterative solving process, which avoids large-scale matrix inversion and significantly reduces the computational burden. Moreover, in view of the sensitivity of the effective regularization parameter to different imaging scenes and array parameters, a fast and effective regularization parameter selection method is developed in this article by employing image correlation (IC) as the evaluation indicator, which contributes to realize the high-precision and fast reconstruction of the target image to the maximum extent. Finally, a large amount of simulation and experimental results demonstrate the superiority of the proposed algorithm in computational efficiency and imaging quality.

engineering, electrical & electronic