改进YOLOv4网络的无人机光通信实时捕获定位技术  

Real-Time Acquisition and Positioning Technology of Unmanned Aerial Vehicle Optical Communication Based on Improved YOLOv4 Network

作　　者：陈廷祚 倪小龙[1] 白素平[1] 于信[1] Chen Tingzuo;Ni Xiaolong;Bai Suping;Yu Xin(School of ElectroOptical Engineering,Changchun University of Science and Technology,Changchun 130022,Jilin,China)

机构地区：[1]长春理工大学光电工程学院,吉林长春130022

出　　处：《激光与光电子学进展》2022年第11期225-234,共10页Laser & Optoelectronics Progress

基　　金：中国博士后科学基金(2017M621179);吉林省科技发展计划(20170521001HJ,20200401054GX);长春理工大学青年基金(XQNJJ2019-01,XJJLG2018-20)。

摘　　要：信标光斑位置检测技术广泛应用于基于视觉的光通信粗对准领域中,而检测算法的优劣直接影响捕获定位的精度。针对基于阈值分割搜寻信标光斑的算法易受背景强光影响的缺陷,建立了基于深度学习算法的无人机光通信实时捕获定位系统。首先,改进了YOLOv4(You only look once,v4)网络,采用能增强浅层特征信息提取的特征图通道拼接方式设计了四个简化模块和一个上采样模块,极大提升了网络的速度。然后,用改进后的网络、原始YOLOv4网络及其简化网络在PASCAL VOC数据集上进行训练。最后,采集和训练信标光斑数据集,在无人机上运行改进YOLOv4网络并输出摄像头视频帧的信标光斑位置。基于比例积分微分算法调节云台进行位置闭环控制,从而实现光通信的实时捕获和定位对准。实验结果表明,改进YOLOv4网络在信标光斑测试集上的精确率为99.6%,召回率为99.8%,在NVIDIA Jetson Xavier NX嵌入式计算机平台上的帧率为42 frame/s,满足无人机光通信实时捕获定位的要求。The beacon spot position detection technology is widely used in the field of visionbased optical communication coarse alignment,and its detection algorithm directly affects the accuracy of acquisition and positioning.Aiming at the defect that the algorithm of searching beacon spot based on threshold segmentation is easy to be affected by background strong light,a realtime acquisition and positioning system of unmanned aerial vehicle optical communication based on deep learning algorithm is established in this paper.First,the YOLOv4(You only look once,v4)network is improved,four simplified modules and one upsampling module are designed using the feature map channel splicing method that can enhance the extraction of shallow feature information,which greatly improves the speed of the network.Then,the improved network,the original YOLOv4 network and its simplified network are trained on PASCAL VOC data set.Finally,collect and train the beacon spot data set,and run the improved YOLOv4 network on the unmanned aerial vehicle to output the beacon spot position of the camera video frame.Based on proportion integration differentiation algorithm,the gimbal is adjusted for position closedloop control,so as to realize realtime acquisition and positioning for optical communication.Experimental results show that the accuracy rate of the improved YOLOv4 network on the beacon spot test set is 99.6%,the recall rate is 99.8%,and the frame rate on the NVIDIA Jetson Xavier NX embedded computer platform is 42 frame/s,which meets the requirements of realtime acquisition and positioning for unmanned aerial vehicle optical communication.

关 键 词：光通信 无人机 YOLOv4网络 比例积分微分算法 捕获定位 

分 类 号：O439[机械工程—光学工程]