摘要
传统基于统计模型的雷达目标检测算法往往仅利用雷达探测回波的能量信息分辨雷达目标,其难以适应密集杂波环境下的雷达探测感知场景,从而显著降低雷达目标探测性能。针对密集杂波背景下的雷达目标检测难题,本文引入随机森林算法,提出一种基于随机森林的雷达目标多维特征检测方法。利用随机森林方法充分融合雷达探测目标回波中时间维、空间维、距离维等多个维度的特征,形成对雷达目标与环境杂波差异的深度刻画,从而实现精准鲁棒的雷达目标检测,显著提升雷达的目标检测性能。最终,基于雷达实测数据对本文所提方法的性能进行验证,实验结果充分验证了本文所提算法的有效性。
Traditional methods based on statistical model only utilize the energy information,which could not adapt to the highly complex and dynamic battlefield environment and result in significantly decrease of Radar detecting performance.Concerning this issue,Random Forest is introduced to take fully advantage of multi-dimensional feature for Radar target detection in this paper.By utilizing the multi-dimensional features of distance,time and space dimension of the target,the difference between radar target and clutter is deeply expressed.Thus it improves target detection performance of Radar under strong clutter significantly.Finally,the performance of the proposed method is verified based on experimental measured data,and the experimental results validate the effectiveness of the proposed method.
作者
欧阳可赛
黄中华
于俊朋
OUYANG Kesai;HUANG Zhonghua;YU Junpeng(Nanjing Research Institute of Electronics Technology,Nanjing 210039,China;Key Laboratory of Inelli Sense Technology,CETC,Nanjing 210039,China;The Unit 93129 of PLA,Beijing 100843,China)
出处
《现代雷达》
CSCD
北大核心
2022年第7期38-42,共5页
Modern Radar
关键词
多维特征
目标检测
随机森林
multi-dimensional feature
target detection
random forest
作者简介
欧阳可赛,男,1990年生,博士,工程师,研究方向为雷达信号处理,机器学习;黄中华,男,1972年生,硕士,高级工程师,研究方向为飞机航电系统、雷达、电子战需求及装备研制管理;于俊朋,男,1984年生,硕士,高级工程师,研究方向为雷达成像与目标识别,雷达信号处理。
