基于机器学习的激光雷达目标自动检测方法研究

Research on Automatic Detection Method of Laser Radar Target Based on Machine Learning

设计激光雷达目标自动检测优化算法,提高雷达目标的对敌准确打击能力。传统的激光雷达目标检测算法采用目标频率时延联合参量估计方法,随着目标运动幅度的增大和外界干扰强度的增强,对目标的检测准确概率下降。提出一种基于机器学习的激光雷达目标自动检测方法。构建了激光雷达目标的回波模型,设计二阶格型陷波器进行信号增强和干扰抑制。采用机器学习算法给信号加以与其匹配的窗函数,计算出激光雷达目标回波信号的矩形包络,在目标可能出现尺度范围内进行机器学习,利用小波变换进行频谱扫描,得到检测统计量和判别函数,以此实现目标自动检测。仿真实验结果表明,采用该算法进行目标检测,准确检测概率提高,具有较好的抗干扰能力。

The optimization design of laser radar automatic target detection, improve the accuracy of radar target enemy combat capability. Traditional laser radar target detection algorithm based on target frequency delay joint param- eter estimation method, with the increase of the target motion range and external interference intensity, the target de- tection accuracy probability decreased. A laser radar target automatic detection method based on machine learning is proposed. The echo model of the laser radar target is built, and the design of the two order lattice notch filter is de- signed. Signal enhancement and interference suppression. Using machine learning algorithms for signal and matching window function to calculate the rectangular envelope of the laser radar target echo signal, the target may appear at a range of scales for machine learning, using wavelet transform spectrum scanning and detection statistics and discriminant function, in order to achieve automatic target detection. The simulation results show that the proposed algorithm can improve the detection probability and improve the ability of anti disturbance.