TY - JOUR
T1 - Radar Imaging with Highly Sparse Range Profiles Based on Data Completion Neural Network
AU - Li, Da
AU - Zhao, Guoqiang
AU - Li, Shiyong
AU - Sun, Houjun
AU - Bao, Jiacheng
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2025
Y1 - 2025
N2 - Data sparsity is the dilemma that high-resolution imaging radar often encounters in practice. Recently, sparse imaging algorithms based on compressive sensing (CS) theory have emerged as commonly adopted methods to address this problem. However, CS-based methods are inherently constrained by the assumption of scene sparsity, which limits further improvements in the quality of image reconstruction. Moreover, these methods are proven ineffective when confronting with highly sparse data scenarios. In this paper, we propose a sparse imaging framework that comprehensively utilizes both the data generation capability of deep neural networks and a non-coherent tomography algorithm with the more relaxed condition on azimuthal sampling. Firstly, a deep neural network (DNN) for data completion is proposed to effectively extrapolate limited-view radar range profiles, generating full-view data with high quality. Subsequently, by incorporating a fast non-coherent imaging algorithm, our method could effectively reconstruct the target image using the completed data from the previous step. Experimental results demonstrate that the proposed approach excels in reconstructing optimal images from highly sparse data (5% degree of sparsity), where other traditional methods even fail. The structure similarity index (SSIM) and image correlation index (ICIM) of reconstructed images are increased by 63.8%
AB - Data sparsity is the dilemma that high-resolution imaging radar often encounters in practice. Recently, sparse imaging algorithms based on compressive sensing (CS) theory have emerged as commonly adopted methods to address this problem. However, CS-based methods are inherently constrained by the assumption of scene sparsity, which limits further improvements in the quality of image reconstruction. Moreover, these methods are proven ineffective when confronting with highly sparse data scenarios. In this paper, we propose a sparse imaging framework that comprehensively utilizes both the data generation capability of deep neural networks and a non-coherent tomography algorithm with the more relaxed condition on azimuthal sampling. Firstly, a deep neural network (DNN) for data completion is proposed to effectively extrapolate limited-view radar range profiles, generating full-view data with high quality. Subsequently, by incorporating a fast non-coherent imaging algorithm, our method could effectively reconstruct the target image using the completed data from the previous step. Experimental results demonstrate that the proposed approach excels in reconstructing optimal images from highly sparse data (5% degree of sparsity), where other traditional methods even fail. The structure similarity index (SSIM) and image correlation index (ICIM) of reconstructed images are increased by 63.8%
UR - http://www.scopus.com/inward/record.url?scp=105002687354&partnerID=8YFLogxK
U2 - 10.1109/TAES.2025.3560257
DO - 10.1109/TAES.2025.3560257
M3 - Article
AN - SCOPUS:105002687354
SN - 0018-9251
JO - IEEE Transactions on Aerospace and Electronic Systems
JF - IEEE Transactions on Aerospace and Electronic Systems
ER -