Research On Micro-motion Feature Extraction And Recognition Technology Of Aerial Target

Micro motion refers to the micro motion other than translation that widely exists in nature.Typical micro-motion includes rotation,vibration and so on.Through the micro motion target echo simulation and signal processing,the relevant parameters reflect ing the target’s own motion and structural characteristics can be obtained,so as to realize the type identification and rendezvous parameter estimation of air targets by radio fuze,and improve the performance of fuze warhead system.In this paper,the echo signal parameter extraction and feature processing of three typical micro motion targets are studied.The main work is as follows:The mathematical model of micro moving target,missile target intersection model and fuze detection field model are established,and the simulation software of micro moving target echo signal calculation and starting characteristics is developed for pulse Doppler Fuze.The occlusion effect in the process of micro motion is studied.According to the principle of computer graphics,two clipping algorithms are used to simulate the occlusion clipping effect between polygons in the process of motion,and the performance of edge by edge clipping algorithm is verified.Based on the circular array antenna,the simulation model of radio fuze detection field and the calculation model of detection area parameters under local irradiation are established.The echo signal simulation is carried out for two kinds of radio fuzes of random two-phase code and pulse compression,two kinds of blocking jamming and towing active jamming,and three typical air micro moving targets of helicopter,propeller aircraft and UAV,the echo signals under different simulation conditions are compared and analyzed.The radar cross-sectional area of the target with micro moving parts is calculated.For the RCS of different targets,a variety of time-frequency analysis methods are used to obtain and analyze the time-frequency diagram.The targets of composite motion types and composite moving parts are analyzed.The synthetic scattering points and local illumination of a small number of panels are simulated,The time-frequency diagram is processed.Finally,image entropy,template matching algorithm and convolution neural network are used to realize target type recognition and intersection parameter estimation.