Research On Radar Target Classification Method Based On Micro-Motion Features

The radar is an important means to detect and identify the target in the battlefield reconnaissance,space surveillance and precision guidance.However,the rapid development of the target feature control technology and the false target image deception jamming technique,such as stealth,camouflage and so on,makes the target detection and identification based on the target shape has been a serious challenge.The further excavation of this unique movement information is a feasible way to deal with this challenge.Its importance has aroused great concern in related fields at home and abroad,and it has become a new hotspot in radar target detection and recognition.Firstly,in the thesis,the Doppler shift expression of the target with micro motion is expressed by the mathematical equation based on the mechanism of micro-Doppler effect.A mathematical model for the micro-Doppler modulation effect of aerial target rotor components was established for helicopter,propeller aircraft and jet aircraft.Secondly,the radar data of helicopter,propeller and jet were analyzed in depth,the micro-Doppler modulation characteristics of the single CPI are observed,then we extract five characteristics of the second-order moments,the waveform entropy in time domain,the amplitude relative amount and the phase relative quantity and the waveform entropy in frequency domain to be the feature vector,and the three kinds aircraft are classified by the support vector machine.The classification results show that the five characteristics of the extraction can effectively classify the three types of aircraft,and the classification accuracy is improved with the increase of the SNR.The characteristics of the micro-Doppler modulation in the combined time-frequency domain are analyzed,and the image entropy feature of the time-frequency map is extracted as the eigenvector to classify the three types of aircraft targets.The classification results are signal-to-noise ratio And the wave entropy and variance of the average time spectrum of the time-frequency spectrum are extracted as the eigenvector.SVM classifier is used to classify the three types of aircraft.The classification results show that the classification accuracy is affected by the signal-to-noise ratio The effect of classification is obviously weakened,and the accuracy of classification increases with the increase of signal-to-noise ratio.This paper uses the GUI graphical interface tool provided by MATLAB to prepare an aircraft radar response and target automatic classification interface,which mainly includes four parameters: radar parameter setting,aircraft parameter setting,target recognition and preservation simulation data.Finally,this paper proposes a method of aircraft target classification based on LPC algorithm.The third-order LPC coefficients are used as the eigenvector to classify the three types of aircraft targets.The results show that under the condition of low pulse repetition frequency and dwell time,Method is still able to achieve better classification results.