High-resolution One-dimensional Range Image Fuze Technology

Using the forward-looking wide-band fuze which follows up the seeker to detect the target can measure the intersection parameter of the target from a long distance,so as to improve the control accuracy of explosion point.In this paper,we will focus on processing high range resolution profile(HRRP)of the target based on the stepped fequency fuze.Four aspects of panoramic range profile splicing,motion compensation,classification and identification,modeling and echo simulation method of typical target will be deeply studied.The main work of this paper is as follows:Based on the high range resolution profile generating of target,an extraction criterion for range profile from arbitrary sampling points is proposed.The method of discarding and maximum value retaining is used to stitch panoramic range profile of simulation target.The simulation results verify the effectiveness of the method.The influence of target velocity on the quality of HRRP is studied.The performance of two range profile velocity compensation algorithms,time-domain cross-correlation method and minimum entropy method,is analyzed by simulation.The velocity compensation and profiling of the simulated moving target are completed by improving a waveform design method of signal in the form of double pulse sequence.Under two kinds of intersection conditions,the high range resolution profile data set of five kinds of simulation targets in a certain intersection angle range is constructed.The one-dimensional Convolution Neural Network is used to classify and recognize the HRRP data set,and good recognition results are obtained.The software of RCS and stepped frequency echo calculation of three target model is developed.The echo calculation and range profile profiling analysis of standard metal sphere and typical target model in static state are carried out.The profiling experiments of static target,UAV target and aircraft target are carried out by using the stepper frequency detector prototype,and the methods described in Chapter 2and Chapter 3 are preliminarily verified.