Research On High Resolution Imaging Method Of Distributed Radar System

In this dissertation,the distributed radar imaging system can utilize space-expanded multiple transmitters and multiple receivers to achieve effective observation of targets.This imaging system integrates the existed radar imaging system,and can establish a unified description of the radar imaging.Based on space spectrum theory,this paper analyzes four typical spatial spectrum filling forms of two typical radar imaging systems i.e.Inverse Synthetic Aperture Radar(ISAR)and Distributed Passive Radar respectively.And it focuses on realizing the high-resolution imaging methods under different spatial spectrum filling forms.The main works and contributions are presented as following:Firstly,the imaging models for ISAR and distributed passive radar system are established.The spatial spectrum imaging theory is used to describe the imaging performance of these two radar systems,and the relationship between the spatial spectrum filling range and the resolution of the imaging system is obtained.The relationship between the degree of undersampling for spatial spectrum and the ill-posed property of the radar imaging inverse problem is further analyzed.Secondly,in the case that the spatial spectrum is filled uniformly and densely,the application of spectral estimation method in distributed radar high resolution imaging is studied.The sensitivity of the spectral estimation method to noise is described.When the imaging scene satisfies sparsity,the received echo matrix has low rank property.Using the low-rank property of the echo matrix,a denoising method based on the matrix low-rank property is proposed to bate the effect of noise on the imaging performance of the spectral estimation method.The research results show that the spectrum estimation method combined with the proposed low rank denoising method can achieve high-resolution imaging robustly.Thirdly,in the case that spatial spectrum is undersampling,a sparse imaging method based on Bayesian model is studied.The proposed Bayesian model can describe the sparse prior information of the target effectively and ensure the accuracy of the solution.At the same time,the imaging result of the scene scattering coefficients can be obtained through iteration,avoiding the tedious process of selecting the regularization parameters.This method can also estimate the parameters of the noise variance accurately and ensure the reconstruction performance under low SNR conditions.Fourthly,a sparse imaging method for expanded target is researched.For an extended target,the sparseness of the scattering points in the imaging scene is reduced and the scattering points are regionally distributed.In addition to the sparsity constraint,we further consider utilizing the structural information of extended target,and propose a variable weight sparse imaging method combining TV regularization constraints method.This method can maintain strong scattering points and the structural distribution of the scattering points simultaneously.For simulation scene imaging,the scattering coefficients are no longer sparse.First,the scattering coefficients of the simulated scene need to be sparsely characterized in the wavelet basis.The wavelet coefficients after the representation have sparsity and tree structure.In this paper,the cascading Bayesian model is used to characterize the sparsity and tree structure of wavelet coefficients,thereby improving the imaging performance of simulation scenes.