Polarization Feature Extraction Of Artifical Target And Inversion Of Meticulous Geometry Structure

High-resolution radar technology such as SAR and ISAR Radar has greatly expanded the application of conventional radar.The introduction of polarization information is not only capable of providing multi-channel data support for radar target recognition work,but also be able to obtain characteristics and physical attribute information of the target.In the traditional field of remote sensing,the polarization of the natural scene has a few achievements,however the polarization information processing of the artificial target is less due to the influence of the data source and so on.And the recognition of the artificial target can be Military reconnaissance,battlefield surveillance and key targets.Based on the electromagnetic scattering characteristics of man-made targets,this paper studies the feature extraction and structural inversion of artificial targets under full polarization conditions.Firstly the research focuses on feature extraction and physical structure inversion of One-dimensional High Resolution Range Profile(HRRP),a generalized orthogonal matching tracking scattering matrix estimation and structural inversion algorithm based on Gai's discs is proposed based on the basic knowledge of Gassian disc and compression sensing.Based on the data description of UAV,with the amplitude and fluctuation characteristics of HRRP taken into account,we establish the novel algorithm for the HRRP problem.The result presents that the classification effect of the algorithm is better than that of the single-polarization under the condition of full polarization.In the work of inversion of one-dimensional target structure,a generalized orthogonal matching tracking scattering matrix estimation and structural inversion algorithm based on Gai's discs are proposed based on the basic knowledge of Gassian disc and compression sensing.The results of the simulated information and the measured data of a certain missile show that the algorithm is efficient and can accurately estimate the position and polarization scattering matrix of the scattering centers,and is capable of accurately extracting the physical structure at the corresponding positions.The ISAR image of the full-polarized two-dimensional UAV is explored by polarization decomposition theory.Three kinds of Pauli,Cameron and Krogager polarization decomposition theory are given,and the mentioned methods are applied to test two kinds of the UAV data,meanwhile the scattering mechanism under different decomposition conditions is discussed and researched.Also on the basis of Cameron decomposition and principal component analysis method,a unique classification and recognition method of principal component analysis based on Cameron decomposition is proposed.The method is exploited to extract the target polarization characteristics on the basis of different types of UAVs and even realize the classification of UAVs at different angles.The 3D structure inversion of radar target is internationally recognized as the difficulty anda foucus.Fusion of polarimetric information and geometric knowledge is adopted to address the issue of the extraction of the geometric size of the three-dimensional objects and the inversion problem with their contour and structure.With the usage of single pitch angle polarization information,three-dimensional structure of the size and the contour can be obtained.This paper mainly focuses on the artificial target of the vehicle,taking the single-polarized electromagnetic simulation data of the vehicle with single pitch angle into account.In the algorithm,backprojection algorithm is applied to extract the parameters of the car,the strong scattering points are extracted by the mechanism of polarization scattering.The method of maximum boundary rectangle is utilized to obtain the length and width of the vehicle.The 3D model of the target vehicle is reconstructed by geometric knowledge and the height of the target is estimated.By comparing with the real value,the errors in each dimension are truly small,and the final scattering picture can accurately reconstruct three-dimensional contour information of the car.