摘要
针对无人机编队数量的检测,文中从信号处理角度,提出了一种基于YOLOv5与脉冲重复间隔(PRI)变换的无人机编队检测算法。首先对无人机旋翼的雷达回波微多普勒信号进行建模与参数化表征,用时频变换方法得到微多普勒时频图,然后对其中的闪烁脉冲用YOLOv5进行检测并得到每个闪烁脉冲的位置信息,最后用PRI变换对闪烁脉冲进行分选,并基于分选结果得到无人机数量的估计。实验结果表明,该方法在信噪比为2 dB时,对4架无人机编队的检测准确率能够达到90%以上。
Aiming at the detection of the number of UAV-formations,an UAV-formations detection algorithm based on YOLOv5 and pulse recurrence interval(PRI) transformation from the perspective of signal processing is proposed.Firstly,the radar echo micro-Doppler signal of the UAV rotor is modeled and parameterized,and then the time-frequency transformation is used to obtain the micro-Doppler time-frequency diagram.The scintillation pulse which in diagram is detected with YOLOv5 and the position of each scintillation pulse is obtained.Finally the scintillation pulses are sorted by PRI transformation and the number of UAVs is estimated based on the sorted result.Experimental results show that the detection accuracy of four UAVs can reach more than 90% when the SNR value is 2 dB.
作者
杨波
孙闽红
仇兆炀
韩文草
YANG Bo;SUN Minhong;QIU Zhaoyang;HAN Wencao(School of Communication Engineering,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China)
出处
《现代雷达》
CSCD
北大核心
2023年第9期53-60,共8页
Modern Radar
基金
国家自然科学基金资助项目(61901149)
国防特色学科发展项目(JCKY2019415D002)。
关键词
无人机
编队数量检测
微多普勒
YOLOv5
脉冲重复间隔变换
unmanned aerial vehicle(UAV)
formation number detection
micro-Doppler
YOLOv5
pulse recurrence interval(PRI)transformation
