Broadband Radar Sea Clutter Statistical Characteristics And Target Detection Technology

With the rapid development of unmanned boats,situational awareness of unmanned boats at sea is very important to navigation safety.Broadband radar is an important equipment to assist ships in safe navigation,and it is more and more widely used on small unmanned boats.The target detection technology of broadband radar can provide important information for situational awareness,collision avoidance decision-making and path planning of unmanned boats to ensure the intelligent navigation of unmanned boats.However,the radar applied to unmanned boats to detect maritime targets needs further research.The main work of this thesis is as follows:(1)Aiming at the problem of complex sea clutter characteristics,this thesis designs and implements a system for analyzing the distribution model of sea clutter amplitude data.Based on the four main distribution models of Rayleigh distribution,lognormal distribution,Weibull distribution and K distribution,parameter estimation and goodness of fit test were carried out on the measured sea clutter data,and the test statistics of the Chi-square test and K-S test were compared and further determines the distribution model with the best fit.This thesis uses radar to conduct sea detection experiments.By analyzing the distribution model of sea clutter data collected under different conditions,it is concluded that the sea clutter distribution model under low grazing angles and the 2-3 class of wind is mainly Weibull distribution;In the case where the low grazing angle,the 3-5 class of wind and there are many targets,measured data mainly Rayleigh distribution;The data distribution model of sea clutter in two spokes with an azimuth interval of 8° under the same range is different.(2)Aiming at the performance problem of the constant false alarm rate detection algorithm under the background of sea clutter,Firstly,this thesis generates simulation data of Rayleigh distribution,Exponential distribution and Weibull distribution.Secondly,by adding targets and increasing the power of simulation clutter data to form uniform clutter and multi-target environment and edge environment of clutter.Thirdly,this thesis use simulation data to compare five constant false alarm rate detection algorithms of CA,GO,SO,OS and VI.The results show that the five algorithms can detect the target better overall.Among them,the detection performance of VI and OS is better,followed by the SO algorithm.In the test in a multi-target environment,it is concluded that the algorithm will appear lose small or weak targets.(3)Target detection problem for radar experimental echo data,the above five algorithms are used for constant false alarm rate detection.Among them,the three constant false alarm rate detection algorithms of CA,GO and SO have the problems of lower threshold and high false alarm rate in the measured data because there is a distance unit with an amplitude value of 0 in the reference unit.For the constant false alarm rate detection algorithm,the elimination of the distance unit with the amplitude value of 0 in the reference unit is added.The detection result shows that this improved method effectively improves the performance of these three algorithms.As for OS-CFAR’s problem of missing alarms in edge clutter and multi-target environment,in this thesis,by adjusting the K-th largest value factor in the algorithm to 0.9times of the reference unit,OS-CFAR has a missing alarm rate in clutter edges and multiple targets,which effectively improves the target detection of OS-CFAR performance.The OSCAFR is used to detect the target in the actual radar data,and the target in the radar echo data and the target azimuth and distance information are obtained.