现代雷达2025，Vol.47Issue(11)：75-82,8.DOI:10.16592/j.cnki.1004-7859.2025070901
基于极化雷达多维特征增强的无人机识别方法
UAV Recognition Method Based on Multidimensional Feature Enhancement of Polarimetric Radar
摘要
Abstract
Radar micro-Doppler features are taken as an important basis for current unmanned aerial vehicle(UAV)model classifi-cation,while the recognition performance is deteriorated sharply when the signal-to-noise ratio(SNR)of radar echoes declines.With the rapid development of radar towards new systems such as multi-frequency,multi-polarization and multi-angle systems,tar-get polarization features are able to provide new ideas for UAV model recognition.In this paper,a UAV recognition method based on multidimensional feature enhancement of polarimetric radar is proposed.Various types of polarization features,including polari-zation cross-correlation parameters,polarization scattering characteristics and polarization invariants,are utilized in this method;meanwhile,strong inter-class discriminative features are extracted through linear discriminant analysis of Doppler spectrum to en-hance multi-class polarization features.Based on measured data,the contribution of multiple polarization features to UAV recogni-tion is studied,and the effectiveness of the recognition method based on multidimensional polarization feature enhancement is veri-fied.Experimental results on a self-built dataset show that the proposed algorithm can effectively improve the UAV recognition rate under low SNR conditions.Compared with the traditional recognition method using original micro-Doppler features,the recogni-tion rates are increased by 13.71%and 9.70%respectively when the main Doppler SNR is 20 dB and 10 dB.关键词
无人机识别/微多普勒特征/极化特征/线性判别分析Key words
unmanned aerial vehicl(UAV)recognition/micro-Doppler features/polarimetric features/linear discriminant analysis(LDA)分类
信息技术与安全科学引用本文复制引用
粟阳健,陈志扬,焦龙翔,王廉钧,王锐..基于极化雷达多维特征增强的无人机识别方法[J].现代雷达,2025,47(11):75-82,8.基金项目
山东省重点研发计划资助项目(重大创新工程)(2025CXGC010209) （重大创新工程）
