Research On High Resolution ISAR Imaging Algorithm Based On Compressed Sensing

Inverse Synthetic Aperture Radar(ISAR)imaging has been widely used in the identification of non-cooperative moving targets such as aircraft and missiles.Therefore,the high resolution of ISAR imaging is important in the identification and feature extraction of non-cooperative moving targets.The traditional high-resolution ISAR imaging technology is limited by the Nyquist sampling theorem,so it has problems such as high sampling rate and large amount of data.Due to the sparse distribution of radar target scattering points,Compressed Sensing(CS)theory can be applied to ISAR imaging,which reduces the difficulty of data acquisition.High-resolution ISAR imaging technology is combined with CS theory,and three types of parameter reconstruction algorithms in CS theory,such as Two Dimension Minimum Smooth l0 Norm algorithm(2D SL0),Gradient Projection Sparse Reconstruction algorithm(GPSR)and Orthogonal Matching Pursuit algorithm(OMP)are studied.Improvements have effectively improved the performance of the algorithm.Three algorithms were improved,and these changes improved the performance of the algorithm.The main work is divided into the following three points:Based on the imaging model of 2D SL0 algorithm,a fast 2D threshold smoothing l0 norm algorithm is proposed to solve the problem of rigid iterative convergence mechanism of conventional 2D SL0 algorithm,which is used to extract strong scattering points in ISAR imaging.The algorithm includes inner and outer loop iterations,and an iterative efficiency index is introduced to evaluate the effectiveness of the inner loop iteration.If the iteration efficiency index is higher than the set threshold,it means that the estimated value of the parameters can be optimized,and the inner loop continues in this round;otherwise,indicating that the parameter estimates have converged in the inner loop of this round,then terminate the inner loop of this round in advance and enter the next inner loop.The experimental results of ISAR imaging show that,compared with the conventional SL0 algorithm,this algorithm can reduce many invalid iterations and significantly reduce the computational complexity of imaging.Based on the conventional GPSR algorithm ISAR imaging model,in view of the large computational load and slow imaging speed of the algorithm,an adaptive weighted GPSR algorithm is proposed for high-resolution ISAR imaging by combining the self-adaptive idea with the GPSR algorithm.In order to speed up the convergence,the algorithm adds penalized weight coefficients to each pixel in the ISAR image,and continuously updates the weights during the gradient descent iteration process.This operation causes small scattering coefficients to converge to zero faster,and large scattering coefficients remain unchanged,speeding up the convergence of the objective function.Through the research on the ISAR imaging model of the OMP algorithm,when the conventional OMP algorithm selects the most relevant atom,each entry of the data is assigned the same weight,and when there is large noise,it will cause serious damage to the entry of the observation matrix,making the selection of atoms wrong,resulting in a serious degradation of the performance of the algorithm.The OMP algorithm based on CIM weighting utilizes the robustness of CIM to assign larger weights to clean data items and smaller weights to severely damaged items,which improves the robustness of the algorithm in ISAR imaging.