Radar Target Recognition And Jamming Identification In Polarimetric Rotation Domain

Radar target recognition plays an important role in space surveillance,earth observation and precision strike.Polarimetric radar can obtain the complete polarimetric scattering matrix of the target,which contains rich scattering information of the target.The polarimetric rotation domain interpretation theory provides a new idea for the deep mining and utilization of target scattering information.This thesis focuses on the target detection and the corner reflector jamming identification in the precision guidance field.A number of polarimetric rotation domain-based method are proposed,including ship detection,man-made target structure recognition and sea corner reflector jamming identification.The effectiveness of the proposed methods is verified by electromagnetic calculation data and real data.The main work and innovations of this thesis are as follows:(1)The mechanism of polarization measurement error on man-made target scattering interpretation is studied.The influence of polarization measurement error on the performance of polarimetric interpretation is analyzed by using the polarization measurement error model and coherent polarimetric target decomposition from three levels of “canonical scattering structure-target component-whole target”.The analytical expressions of the polarimetric target decomposition components and measurement error sources of the canonical scatterers such as trihedral,dihedral,dipole and helix are derived,and the mechanism of the influence of polarization measurement error on the performance of polarimetric interpretation is revealed,which provides a basis for feature selection in the subsequent target structure recognition method.Based on the electromagnetic calculation data of Slicy model and the darkroom measured data of "pioneer" UAV,the quantitative analysis is completed,and the index requirements of polarization isolation and polarization channel imbalance that need to be met in radar target recognition are given.(2)A ship target detection method with the optimal polarimetric rotation domain features and support vector machine(SVM)classifier is proposed.The core idea of this method is to mine and fuse the information hidden in target scattering diversity using polarimetric coherence/correlation pattern and other polarimetric rotation domain interpretation tools for ship detection.Firstly,four independent polarimetric coherence and polarimetric correlation patterns are constructed.Secondly,the Relief method is used to select the optimal polarimetric rotation domain feature set with strong discrimination between ship and sea clutter.Finally,the ship detection issue is transformed into the problem of ship and sea clutter classification,and the ship detection is realized by SVM classifier.The superiority of the proposed method is verified by using polarimetric synthetic aperture radar(SAR)data.The figure of merit(Fo M)is better than 95% in inshore dense ship area.(3)A man-made target structure recognition method with polarimetric rotation domain feature coding is proposed.The core idea is that the polarimetric correlation pattern of canonical man-made scattering structure has obvious different characteristics.Firstly,the difference of polarimetric rotation domain characteristics of canonical manmade scattering structures and the potential of polarimetric correlation pattern in distinguishing different canonical man-made scattering structures are revealed.Secondly,the polarimetric roll-invariant feature set with strong separability is extracted and optimized from the polarimetric correlation pattern,and the feature coding vector of canonical man-made scattering structure is constructed.Finally,the target structure recognition is achieved based on the similarity measure between target and canonical man-made scattering structure features.Experimental studies are carried out by using the electromagnetic calculation datasets of canonical man-made scattering structures and UAV target and real spaceborne polarimetric SAR data of a ship target.Compared with the famous Cameron decomposition,the proposed method has better recognition performance for oriented structures and is suitable for different system measurement errors.(4)The polarimetric rotation domain characteristics of the new type of sea corner reflector are studied,and the variation law of polarimetric rotation domain characteristics of the corner reflector under omnidirectional angle is revealed.Then,the polarimetric rotation domain characteristics difference between the ship target and the corner reflector is analyzed.Based on this,a highly identifiable polarimetric identification feature is designed,and a new identification method for the corner reflector jamming in polarimetric rotation domain is proposed,which effectively realizes the identification of the corner reflector jamming.In addition,combined with the real polarimetric SAR data and the electromagnetic calculation data of canonical corner reflectors,the polarimetric radar simulation dataset with corner reflector jamming is generated through signal-level imaging superposition simulation,and the superiority of the proposed method is verified by using the dataset.The accuracy rate of corner reflector jamming identification is better than 98%.