Research On Radar Imaging Method Based On Block Sparse Variational Bayes

The resolution of radar imaging is determined by the bandwidth of the signal and the size of the coherent accumulation angle,but it is not easy to increase the bandwidth of the signal and the target observation time.Multi-radar data fusion has become an effective means to improve the imaging resolution.The signal reconstruction algorithm is used to estimate the parameters of the target scattering center model,and the radar echo data of some frequency bands and angles of view are recovered to obtain high resolution images.In this paper,the application of block-sparse variational bayesian in radar imaging is studied.Firstly,the principle of imaging is analyzed,and the data of multiple radar echo are represented sparsely,and the data of the blank frequency band and azimuth are reconstructed to improve the imaging resolution.Then we use this method to achieve high quality imaging for maneuvering targets with missing azimuth data.The main tasks completed are as follows:Firstly,the relationship between electromagnetic scattering and imaging is studied.The theory of target scattering center in high frequency region is analyzed.The principle of radar imaging is introduced,and the theory of compressed sensing is briefly described.Secondly,the multi-radar data fusion imaging based on block-sparse variational bayesian is studied.One-dimensional sparse representation matrix is constructed for the multi-band radar echo from the same view,and the frequency data of the blank segment is reconstructed.The imaging resolution is improved at the distance direction and the anti noise performance of the algorithm is verified.The two-dimensional sparse representation matrix of multi view and multi band radar echoes is constructed,the frequency and azimuth data of the blank segment are reconstructed,and the imaging resolution is improved in distance and azimuth.Finally,the maneuvering target imaging based on block-sparse variational bayesian is studied.The error caused by the translation of the target is analyzed.The distance alignment and phase compensation methods are analyzed.The accuracy of the motion compensation method is verified by simulation.For the maneuvering target imaging,the imaging method based on time-frequency analysis is given.A sparse model of the maneuvering target echo is established.When the target azimuth data is missing at random,the clear image of the target is obtained by the block-sparse variational bayesian algorithm.