CS-based high-resolution ISAR imaging with adaptive sparse basis

The theory of compressed sensing (CS) indicates that the optimal reconstruction of an unknown ISAR signal with intrinsical sparsity can be achieved by solving a sparsity-driven optimization problem. Due to this property, a novel method to construct sparse basis is presented for ISAR image generation. We utilize discrete chirp fourier transform (dcft) and CLEAN to decompose the signal of interest into multiple LFM or chirp signals. The proposed method can generate adaptive sparse basis according to the frequencies and chirp rates of these chirp signals. Our proposed method which uses much fewer measured data can get almost the same image compared with the uncompressed conventional imaging algorithm and our proposal outperforms the CS-based method with standard Fourier basis. Both simulated and real experimental results can demonstrate the effectiveness and feasibility of the method.