Research On Low-resolution Radar Target Classification Based On Micro-motion Characteristics

Target recognition is one of the basic capabilities necessary for modern radar.Considering that the current air defense radars are mostly low-resolution radars,this thesis takes three types of aircraft targets: helicopter,propeller and jet airplane,as classification objects,and studies the extraction method and optimization technology of micro-motion characteristics based on the data of low-resolution radar.Also through the simulation experiment to verify the character representation ability and the improvement effect of the optimization technology.Firstly,the existing aircraft target rotor echo model is introduced.For the problem that the model cannot determine the amplitude of the fuselage component,a method of estimating the energy ratio between the fuselage and the rotor of measured data by CLEAN algorithm is proposed.The analysis of the micro-motion characteristics of the simulated data shows that the method of target classification by estimating the spectral line spacing and number in high resolution mode is not suitable for the low resolution mode.Then,three types of micro-motion feature extraction methods are analyzed from three angles: echo waveform,echo component and time-frequency characteristics,including frequency domain waveform entropy,waveform relative quantity,eigenvalue spectrum feature and time-frequency diagram feature.The results of classification simulation experiments show that the three types of features can achieve an average classification accuracy of 87% when the echo signal-to-noise ratio is 15 dB.Then,for the problem that the micro-motion features are greatly affected by noise and fuselage component,three optimization methods of feature extraction are introduced: fuselage component separation based on CLEAN algorithm,micro-motion component enhancement based on BOX-COX,noise reduction based on sparse reconstruction.The three methods improve the expression of the micro-motion characteristics in the echo from the perspective of separating the fuselage component,enhancing the micro-motion component,and reducing the noise.The results of comparative simulation experiments show that all three methods can improve the average classification accuracy.Finally,based on the above research,the radar target automatic classification scheme is given to provide reference for its engineering implementation.