DEEP UNFOLDING FOR SPARSE APERTURE ISAR IMAGING VIA COMPRESSED SENSING: CIST+

The issue of inverse synthetic aperture radar (ISAR) imaging under sparse aperture (SA) conditions using compressed sensing (CS) techniques has garnered substantial attention. In this research, we amalgamate the explicability offered by the deep unfolding network (DUN) with the facile compression of the high-frequency component of the signal in the image domain. This fusion aims to unfold the iterative soft-thresholding algorithm (ISTA) into a trainable network model termed CIST+. Through end-to-end training, the model demonstrates its prowess in achieving robust reconstruction of low-compression-rate ISAR imaging even amidst noise interference, while also exhibiting exceptional operational efficiency. Our findings from simulations and measured data experiments demonstrate the efficacy and potential of the model.