Abstract:Sparsity-driven inverse synthetic aperture radar (ISAR) imaging can obtain a high-resolution image of non-cooperative targets using compressive sensing (CS)-based methods. However, traditional CS-based imaging methods generally require extensive computational complexity, and some fixed parameters need to be manually tuned to be optimal for better imaging quality. Iterative-thresholding methods are typical CS-based algorithms with representative framework and efficient performance. This paper proposes a general structure of hierarchical channel-mixed framework (HCMF) for sparsity-driven ISAR imaging. Specifically, HCMF is constructed by unrolling the iterative-thresholding methods, and each phase of HCMF is mapped from one iteration in traditional iterative-thresholding methods, maintaining the hierarchical framework. HCMF also involves mixed channels that flow separately and together at different processes, to process complex data properly. Besides, the parameters in each phase are set to be learnable and self-adapted optimal, contributing to greater robustness. Without loss of generality, we take approximate message-passing (AMP) for demonstration, dubbed as HCM-AMP-Net (HCM-AMPN). HCM-AMPN has the almost same computational complexity as traditional AMP, but the high-parallelization of deep network accelerates computational efficiency dramatically. Experimental results based on simulated and measured data demonstrate that HCM-AMPN has higher efficiency and superior performance compared with some state-of-the-art methods. In our experiments, HCM-AMPN is more than 100 times faster than traditional AMP, and tens of times faster than other network-based methods.

Scattering region weighting compressive sensing ISAR imaging

ISAR 2-D Imaging under Low SNR Based on Improved Compressive Sensing

Improved ISAR Imaging by Exploiting the Local Structures of the Target Scene.

Wide Angle SAR Subaperture Imaging Based on Modified Compressive Sensing

An Imaging Method Based on Compressive Sensing for Sparse Aperture of SAR

An Iterative Adaptive Reweighted Norm Minimization Sparsity Autofocus Algorithm Via Bayesian Recovery for Array SAR Imaging.

Bistatic ISAR Radar Imaging Using Missing Data Based on Compressed Sensing.

Enhanced Isar Imaging by Exploiting the Continuity of the Target Scene

Compressive Sensing Based Multilevel Fast Multipole Acceleration for Fast Scattering Center Extraction and ISAR Imaging.

Deep learning network with new weighting strategy for ISAR image enhancement

Method for Achieving High Resolution ISAR Imaging with Short Aperture Data Via Compressed Sensing

Achieving High-Quality Three-Dimensional InISAR Imageries of Maneuvering Target Via Super-Resolution ISAR Imaging by Exploiting Sparseness.

Multichannel And Wide-Angle Sar Imaging Based On Compressed Sensing

Multifrequency Compressed Sensing for 2-D Near-Field Synthetic Aperture Radar Image Reconstruction

SPARSE RECONSTRUCTION FOR SAR IMAGING BASED ON COMPRESSED SENSING

Adaptive Sparse Recovery by Parametric Weighted L Minimization for ISAR Imaging of Uniformly Rotating Targets

An Iterative Weighing Algorithm for Image Reconstruction in Compressive Sensing

Sparsity-Driven ISAR Imaging Via Hierarchical Channel-Mixed Framework

An Autofocus Imaging Method for ISAR Based on Bayesian Compressive Sensing

Linear Array Sar Imaging Via Compressed Sensing

A Computational Efficient 2-D Block-Sparse ISAR Imaging Method Based on PCSBL-GAMP-Net