Research On The Method Of SAR Images Super-resolution Based On Match Pursuit Algorithm

SAR images super-resolution technique is one of the main aim of developing SAR technology. Improving and updating hardware can increase the resolution of SAR images effectively. Due to the limitations of high cost, long development cycles and techniques level for hardware, Investigating SAR images super-resolution technique by signal processing method has important theoretical and practical significance.Relying on mathematical theories and methods, this paper investigates a new method of SAR images super-resolution processing based on the sparse representation and the point-scattering model.The contributions of this paper are listed as follows:From the view of sparse representation of the signal, this paper discusses the sparsity of the SAR image's point-scattering model, connects the sparsity of SAR images and SAR images super-resolution processing, provides the foundation for extending the technique of sparse representation to SAR images super-resolution processing.By constructing a compacted dictionary in sub-pixel, the precision of parameter estimation is improved and the performance of super-resolution processing is enhanced. A modified iterative algorithm based on Match Pursuit is designed for overcoming the "short sight" of the Match Pursuit. Simulation experiments and computational results of measured data demonstrate that the method can suppress the main lobe and increase the resolution efficiently, at the same time the side lobe is also suppressed.The performance of the method is analysed by a series of simulation experiments based on the sparsity of SAR images, the noise and the dictionary of Match Pursuit. Whole sparsity and local sparsity are proposed aiming at the sparsity of SAR images.