Noise-robust interferometric ISAR imaging of 3-D maneuvering motion targets with fine image registration

Low signal-to-noise ratio (SNR) and three-dimensional (3-D) maneuvering motion of targets present enormous challenges to the interferometric inverse synthetic aperture radar (InISAR) imaging. In this paper, we propose a noise-robust InISAR imaging algorithm to acquire high-precision InISAR images of 3-D maneuvering motion targets at low SNR. With respect to time-variant wave path difference (TVWPD) and spatial-variant wave path difference (SVWPD), a novel algorithm, named non-coherent fusion image registration algorithm (NCFIR), is developed to realize fine image registration. By implementing the non-coherent fusion processing on the multi-look ISAR images, it improves the image SNR and therefore makes possible the joint compensation of TVWPD and SVWPD at low SNR. In addition, underpinned by the Bayesian compressive sensing (BCS) theory, a joint multi-channel BCS ISAR imaging algorithm (JMC-BCS) is also proposed. Its capability for suppressing the strong noise and high side lobes of ISAR imaging results effectively improves the SNRs of imaging results. Through iterative processing of JMC-BCS and NCFIR, high-quality 3-D InISAR images of 3-D maneuvering motion targets at low SNR can be obtained. The proposed algorithm improves the SNR of image twice through JMC-BCS and NCFIR, rendering it more robust against noise. Extensive experimental results from both simulated and real data corroborate the effectiveness and robustness of the proposed algorithm which outperforms other available InISAR imaging approaches at low SNR. (C) 2022 Published by Elsevier B.V.