中文摘要：

In compressive sensing(CS) based inverse synthetic aperture radar(ISAR) imaging approaches, the quality of final image significantly depends on the number of measurements and the noise level. In this paper, we propose an improved version of CSbased method for inverse synthetic aperture radar(ISAR) imaging. Different from the traditional l1 norm based CS ISAR imaging method, our method explores the use of Gini index to measure the sparsity of ISAR images to improve the imaging quality. Instead of simultaneous perturbation stochastic approximation(SPSA), we use weighted l1 norm as the surrogate functional and successfully develop an iteratively re-weighted algorithm to reconstruct ISAR images from compressed echo samples. Experimental results show that our approach significantly reduces the number of measurements needed for exact reconstruction and effectively suppresses the noise. Both the peak sidelobe ratio(PSLR) and the reconstruction relative error(RE) indicate that the proposed method outperforms the l1 norm based method.

英文摘要：

In compressive sensing(CS) based inverse synthetic aperture radar(ISAR) imaging approaches, the quality of final image significantly depends on the number of measurements and the noise level. In this paper, we propose an improved version of CSbased method for inverse synthetic aperture radar(ISAR) imaging. Different from the traditional l1 norm based CS ISAR imaging method, our method explores the use of Gini index to measure the sparsity of ISAR images to improve the imaging quality. Instead of simultaneous perturbation stochastic approximation(SPSA), we use weighted l1 norm as the surrogate functional and successfully develop an iteratively re-weighted algorithm to reconstruct ISAR images from compressed echo samples. Experimental results show that our approach significantly reduces the number of measurements needed for exact reconstruction and effectively suppresses the noise. Both the peak sidelobe ratio(PSLR) and the reconstruction relative error(RE) indicate that the proposed method outperforms the l1 norm based method.