Research Of Radar Imaging Methods Based On Sparsity

Radar imaging has developed rapidly in the fields of national economy and national defence during recent years.With the rapid change of demand,traditional radar imaging cannot meet the requirement of imaging with incomplete data.Radar imaging belongs to special signal reconstruction from the viewpoint of signal processing.To be an important development trend of radar imaging,sparse reconstruction algorithm can realize the incomplete data imaging under sparse assumption.Sparse based radar imaging is of important meaning for its advantages compared with traditional methods.In our research we introduced radar imaging methods of reconstruction theory.System model,reconstruct methods and imaging methods are studied.According to different sparsity,radar imaging applications have been researched from the view of general sparse,block sparse and joint sparse.A series of theory and algorithms have been proposed and simulation experiments have been conducted to validate the effectiveness.Based on the electromagnetic scattering theory,radar scene can be modeled according to a few strong scattering points for its potential sparsity.We begin from the radar echo model and derived an approximate observation model which is suitable for a variety of radar imaging modes and a joint sparse model which is suitable for multi-channel scanning radar imaging beam sharpening.The research builds the foundation of radar imaging under general sparse,block sparse and joint sparse.Firstly,aiming at low convergence speed of present general sparsity radar compressed sensing imaging methods,an AMP based radar imaging method is proposed based on approximate observation model.With same single iterative computation unchanged,the convergence rate has been accelerated and imaging efficiency has been improved.For the difficult issue to add complex structured prior information in present general sparsity radar compressed sensing imaging methods,we proposed an OMP based compressed sensing imaging method,which made the greedy algorithm possible to be utlized in compressed sensing radar imaging.It laid the foundation for complex structured prior information based radar imaging.Secondly,since the existing imperfect block sparse cannot handle non-uniform block signal,a non-uniform block sparse theory has been proposed and block mutual coherence hasbeen defined.Also the rule of uncertain has been given to confirm the uniqueness condition of sparse representation,the performance has been evaluated.By expanding BOMP algorithm to non-uniform block sparse signal reconstruction,the algorithm performance can be enhanced and applied range can be widened.Block sparse based ABOMP compressed sensing radar imaging algorithm has been proposed to realize block sparse radar imaging.Traditional ABOMP algorithm has low efficiency for its least square algorithm,gradient descent based atomic orthogonalization has been put forward to improve the efficiency.Finally,joint sparse based multi-channel beam sharpening method has been studied.Aiming at the problem that present methods of beam sharpening cannot use the multi-channel characteristics and sparse feature simultaneously,an MCIST multi-channel beam sharpening method based on joint sparsity that based on the combination with the parameter estimation theory is presented.Thanks to the joint sparsity which simultaneously uses multi-channel characteristics and sparse features,the new method achived multi-channel beam sharpening meanwhile improving the beam sharpening performance.