Research On Dual-polarization Radar Joint High-resolution Imaging And Target Recognition

Radar high-resolution imaging technology is a basic content in the research of space target detection and recognition.Multi-radar data fusion technology based on compressed sensing theory has become an effective means to improve imaging resolution.The target reconstruction center parameters are estimated by signal reconstruction algorithm.Data with missing observation frequencies and viewing angles are constructed to achieve high resolution imaging of radar.Polarization information is another important information that can be utilized in addition to the target time,frequency and space.In recent years,research on target scattering characteristics has made in-depth progress.In this paper,based on the application of the compressed sensing reconstruction algorithm of dual polarization measurement system in radar imaging,the combination of polarization combined sparsity theory and data fusion technology is superior to traditional monopole in the same SNR.Then the method is applied to the data preprocessing process of subsequent target recognition and combined with the multi-classifier joint recognition network to improve the recognition accuracy of the target.This paper introduces the dual-polarization combined high-resolution imaging technology and its application in target recognition through the following parts:The first part introduces the application background of radar high resolution imaging and target recognition.Secondly,the theoretical basis of multi-radar data fusion technology is introduced,including signal sparse representation theory,target scattering center theory and basic theory of compressed sensing.The second part studies the radar high resolution imaging technology based on the dual polarization joint parameter estimation method.Firstly,the flow based on the dual-polarization joint parameter estimation algorithm is given from the single-polarized Bayesian parameter estimation algorithm combined with the polarization joint sparsity theory.Secondly,based on the one-dimensional and two-dimensional echo signal sparse representation model of radar echo signals,the principle of multi-radar data fusion high-resolution imaging technology based on the principle of compressed sensing is introduced.Finally,the experimental results show that the method has good anti-noise performance.The third part studies the application of dual-polarization combined reconstruction imaging technology in target recognition.Firstly,the dual-polarization joint parameter estimation method is used to reconstruct the target echo signal to be reconstructed to improve the influence of the noise signal on the target echo.Secondly,the feature fusion selection technology based on principal component analysis is studied,which makes the features used for target recognition more streamlined and effective.Finally,the joint recognition network architecture based on multi-classifier is studied to improve the generalization ability of single classifier.By constructing the joint classification accuracy loss function,the best decision weight is assigned to each classifier and finally improved by joint decision.