Research On Detection And Recognition Of Non-Cooperative Targets In Space Based On Laser Micro-Doppler Effect

Accurate measurement of the relative motion parameters of non-cooperative targets in a large range from far to near is a difficult problem in the field of autonomous rendezvous and docking and in-orbit maintenance.At present,the measurement methods based on machine vision have shortcomings such as short detection range and poor real-time performance.The motion of the space target itself,such as precression,spin,roll and vibration,produces Doppler modulated sideband on the detection signal,which provides the possibility of extracting the relative motion parameters of the target based on the Doppler sideband,and is expected to reverse the velocity,angular velocity,attitude Angle and size of the non-cooperative target from it.The coherent detection lidar in near-infrared frequency band has natural advantages in detection sensitivity and accuracy.Therefore,the measurement technology of multi-dimensional motion parameters of space non-cooperative targets based on Micro-Doppler lidar has become a research hotspot.In this paper,the microwave Micro-Doppler echo model of Cube Sat is established based on the "point scattering model" which is relatively mature in microwave radar,and the simulation results of echo signal are given.According to the difference of target scattering characteristics between laser frequency band and microwave frequency band,the "surface scattering model" of laser frequency band is established,and on this basis,the laser Micro-Doppler echo model of Cube Sat is established,and the simulation results of echo signal are given.Secondly,according to the characteristics of the echo signal simulation time-varying non-stationary,the short-time Fourier transform in time-frequency analysis theory method to calculate the cube satellite microwave under typical parameter of micro-doppler signal spectrum and cube satellite laser micro-doppler signal spectrum,summarizes the time-frequency characteristics of target under different simulation parameters change rule;In view of the characteristics of low SNR of actual signals,a time spectrum image preprocessing method based on image enhancement technology is proposed,so that the low frequency details of the time frequency map are preserved,the edge noise is suppressed,and the internal texture is smooth,thus laying a foundation for multi-dimensional parameter extraction based on time frequency characteristics.Then,a Cube Sat Relative Attitude Measurement Algorithm(CRAMA)based on timefrequency characteristics is proposed,and an adaptive threshold processing method of time-frequency graph is proposed based on the idea of least squares.The CRAMA algorithm does not need the target size as the prior information to solve the three parameter indexes of the Cube Sat target: spin frequency,attitude Angle and side length.The simulation results show that when the SNR is high,the accuracy of spin frequency extraction is 100%,the absolute error of attitude Angle extraction is controlled within 2°,the relative error is less than 2%,and the absolute error of cube side length extraction is less than 0.003 m,the relative error is less than 6%.Finally,an all-fiber laser coherent detection system with a wavelength of 1064 nm is designed and built.The simulated Cube Sat targets with different spin frequencies are detected,and the extraction results of time-frequency features,spin frequency and attitude Angle parameters of the simulated Cube Sat targets are given.The experimental results show that the absolute error and relative error of the two targets are less than0.2Hz and less than 7.5%.The absolute error of the attitude Angle extraction of the cube target is less than 5°,and the absolute error of the attitude Angle extraction of the cube target with rectangular sailboard is less than 10°.