Research On Technologies Of Radar CFAR And Revolving Target Detection

As one of the important developing directions of modern radar technology,Radar target detection and recognition has become an important component in the weapon system of countries in future.Therein,constant false-alarm rate (CFAR) testing has become one of the key technologies in radar signal processing.It will have a direct impact on radar detection performance.Target recognition technology takes the initiative in battle by the correct and effective identification of target.This paper focuses on the CFAR detecting method,feature extraction and the recognition technology of helicopter goals and spin targets.Based on the introduction of radar signal detection principle and relevant theory, this paper focus on researching the CFAR detection method.Summarizing the structure and principle of several typical parameter model CFAR detector under the background of rayleigh,webb's and logarithmic - normal clutter.Analyzing comparatively the detection performance of several kinds of single parameter and double parameters CFAR detector under the background of webb's and logarithmic - normal clutter.In addition,Aiming at the non-gaussian environment,the performance of traditional CFAR detection method based on sliding window self-adaptive threshold will decline,this paper proposed a CFAR detection methods based on goodness-of-fit (GOF) test.Simulation results show that background distribution characteristics have a little effect on this method,and it has strong adaptability to non-gaussian environment. Studying the recognition problem of revolving targets.Aiming at the problem of helicopter recognition,this paper put forward identifying the helicopter goals by periodic signals autocorrelation testing method. Aiming at the problem of spin targets, this paper put forward a space target recognition algorithm based on range profile. Extracting target radial length direct from1D range profile,then the revolving target can be recognized by the standard deviation of radial length sequence.Aiming at the low SNR cases,the recognition rate of above method is lower,this paper proposed a method which use wavelet de-noising and RELAX algorithm to extract target radial length, Simulation results show that this method can obtain high recognition rate in low SNR.