基于改进YOLOv5的输电线路多目标检测  

Multi-target Detection of Transmission Lines Based on Improved YOLOv5

作　　者：汤浩威 姚军财 姚聪颖 孙颖 裴星懿 宋春晓 TANG Hao-wei;YAO Jun-cai;YAO Cong-ying;SUN Ying;PEI Xing-yi;SONG Chun-xiao(School of Electrical Engineering,Nanjing Institute of Technology,Nanjing 211167,China;School of Computer Engineering,Nanjing Institute of Technology,Nanjing 211167,China)

机构地区：[1]南京工程学院电力工程学院,江苏南京211167 [2]南京工程学院计算机工程学院,江苏南京211167

出　　处：《计算机与现代化》2023年第2期78-82,共5页Computer and Modernization

基　　金：国家自然科学基金资助项目(61301237);江苏省自然科学基金资助面上项目(BK20201468);南京工程学院高层次引进人才基金资助项目(YKJ201981)。

摘　　要：针对当前目标检测网络层数加深、参数量和计算量加大,造成实时性差等问题,为了实现对输电线路部件的识别与检测,提出一种基于改进YOLOv5的输电线路多目标检测算法。首先,使用ShuffleNetv2结构作为网络特征提取的主干结构,减少网络的参数量;然后,将PANet网络中的BottleneckCSP改为Light_CSP模块,加快特征融合的速度;其次,使用CIoU loss、DIoU-NMS方法减少预测框的位置损失和漏检问题。最后,为了验证所提算法的有效性,利用输电线路图像数据集进行训练与测试。结果表明,改进YOLOv5的参数量为7.5×106,浮点计算量为10.9,平均精度达到了87.5%,FPS达到69.2,能够满足输电线路部件检测的精度、轻量化与实时性要求。In order to realize the identification and detection of transmission line components,a multi-target detection algorithm for transmission lines based on improved YOLOv5 is proposed for the current problems of deepening the number of target detec⁃tion network layers,increasing the number of parameters and computation,resulting in poor real-time performance.Firstly,the number of parameters in the network was reduced by using the shuffleNetv2 structure as the backbone structure for network fea⁃ture extraction.Secondly,the BottleneckCSP in the PANet network is changed to a Light_CSP module to speed up feature fusion.Thirdly,the CIoU loss,DIoU-NMS method is used to reduce the loss of position of the prediction frame and the problem of missed detection.Finally,in order to verify the effectiveness of the proposed algorithm,a transmission line image dataset was used for training and testing The results show that the improved YOLOv5 has a parametric count of 7.5×106,a floating point com⁃putation of 10.9,an average accuracy of 87.5%and an FPS of 69.2,which meets the requirements for accuracy,lightness and real-time inspection of transmission line components.

关 键 词：智能巡检 目标检测 YOLOv5 输电线路 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]