HRRP Feature Reconstruction Based On Sinuular Function Analysis Model

The high resolution range profile(HRRP)is the superposition of the target backscattered echoes received by the wideband radar in the distance unit.The HRRP provides the distribution of the target scattering intensity along the radar line of sight.It has the advantages of easy acquisition and simple processing.Therefore,HRRP be-comes a hot research area in the field of radar automatic target recognition.First of all,the properties of range profiles,feature extraction and target classifi-cation methods for HRRP are introduced in this thesis.On this basis,a singular point method is used to extract the scattering center position feature to solve the aspect sen-sitivity problem.Since the singular points can accurately characterize the scattering center of the HRRP,the singular point of HRRP is extracted to construct the scattering center position feature.In the simulation experiments,the recognition performance of the scattering center position feature,singular point and other features representing the scattering center position are compared.The simulation results show that the scattering center position is more aspect robust than the scattering center intensity feature.Com-pared with other scattering center extraction algorithms,the singular points can obtain the strong scattering center of HRRP more accurately.To deal with the problem of information missing and translation sensitivity,the idea of reconstructing images originally proposed for the magnetic resonance images processing is introduced to the range profiles recognition community.Because the sin-gularity function analysis model can reconstruct complete signal from partial spectral data,the original HRRP and the FFT amplitude can be reconstructed to get more com-prehensive information.In the simulation experiments,the BP neural networks and the improved algorithms are selected as classifiers to evaluate the recognition performance of original range profile and FFT amplitude reconstructed feature.The simulation re-sults show that the reconstruction algorithm is effective to improve the target recogni-tion performance.