MDI训练样本集构建对雷达探测旋翼无人机分类的影响分析  

Analysis of Impacts of MDI Training Samples Sets Construction on Classification of Rotor UAVs Detected by Radar

作　　者：吴仁彪[1] 黄诚 王晓亮[1] 何炜琨[1] 刘闪亮 WU Renbiao;HUANG Cheng;WANG Xiaoliang;HE Weikun;LIU Shanliang(Tianjin Key Laboratory for Advanced Signal Processing,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学天津市智能信号与图像处理重点实验室,天津300300

出　　处：《信号处理》2021年第6期1017-1033,共17页Journal of Signal Processing

基　　金：天津市研究生科研创新项目(2019YJSS071)。

摘　　要：利用卷积神经网络对目标微多普勒特征进行深度学习是目前雷达探测无人机分类的重要手段。实际应用中,无人机参数如叶片转速、叶片长度、叶片初始相位、无人机方位角、无人机俯仰角、无人机径向速度等参数变化大,导致训练样本变化大。该文分析训练样本集对旋翼无人机分类结果的影响。首先建立单旋翼无人直升机、四旋翼无人机和六旋翼无人机雷达回波仿真模型。然后对其进行微多普勒特征分析提取,构建多种不同情况下的合并多普勒图像(Merged Doppler Images,MDI)训练样本集。最后利用GoogLeNet(Inception v1)得到不同情况下的无人机分类结果,分析训练样本集中样本数量、无人机单一参数变化、样本参数涵盖完整性以及无人机参数采样间隔对分类准确率的影响。实验结果表明:训练样本集的差异可能对分类准确率产生显著影响。It is an important method for classifying UAVs detected by radar using convolutional neural networks to perform deep learning on targets’micro-Doppler features.Actually,parameters of UAVs such as blade rotation speed,blade length,blade initial phase,UAVs’azimuth,UAVs’pitch angle,and UAVs’radial velocity,etc.vary greatly,which leads to large variation in training samples sets.In this paper,impacts of training samples sets on rotor UAVs’classification results are analyzed.Firstly,simulated radar echoes models of helicopters,quadrotors and hexarotors are established.Then micro-Doppler features analysis and extraction are carried out,and Merged Doppler Images(MDI)training samples sets are constructed in many different situations.Finally,GoogLeNet(Inception v1)is used to obtain UAVs’classification results in different situations.Impacts of sample quantity,variation of UAVs’single parameter,completeness of sample parameters coverage and sampling intervals of UAVs’parameters of training sets on the classification accuracy are analyzed.The experiment results show that the difference in MDI training sets may have significant impacts on UAVs’classification accuracy.

关 键 词：雷达目标分类 无人机探测 卷积神经网络 微多普勒特征 GoogLeNet 

分 类 号：TN957[电子电信—信号与信息处理]