A Sparse Imaging Reconstruction Algorithm for Near-Field SAR Based on Rényi Entropy Regularization

Near‑field synthetic aperture radar imaging holds substantial promise for applications in counterterrorism security and nondestructive testing. However, most existing imaging algorithms demand large data volumes, which hinders their practical deployment. Although compressed sensing based sparse imaging can reduce sampling requirements, conventional sparse reconstruction methods still suffer from a prohibitive sensing‑matrix size, disruption of patch‑wise sparsity structures, and difficulty in adaptively tuning regularization parameters. To address these issues, we propose a sparse imaging algorithm with Rényi‑entropy regularization. Building on a two‑dimensional sparse imaging model that markedly reduces the dimensionality of the sensing matrix, we introduce the Rényi‑entropy function as a nonconvex, energy‑concentrating regularizer that preserves target geometry. In addition, a modified Barzilai–Borwein scheme is employed to adaptively optimize the regularization parameter, thereby improving algorithmic robustness. To validate the effectiveness of the proposed approach, we developed a millimeter‑wave radar data‑acquisition platform and collected real measurement data. Experimental results show that, for imaging scenarios with non‑concealed and concealed targets, the proposed method achieves high‑resolution reconstructions using only 25% and 56.25% of the fully sampled data, respectively, providing strong technical support for high‑resolution near‑field radar imaging in complex environments.