摘要
绝缘子是高压输电线路的重要组成部分,对维护输电线路的稳定和保证输电网的正常运行具有重要意义。绝缘子一旦出现故障,将会造成严重的输电故障和经济损失。为此,文章提出了一种改进YOLOv3的高压输变电线路绝缘子检测方法,该方法首先使用特征提取能力更强的VoVNet作为主干网络,其次提出了一种新型的特征增强模块,可以有效提升浅层特征图的语义信息和感受野。通过对比实验表明,文章提出的方法mAP达到98.01%,相较于原始YOLOv3算法提升了6.07%,在保证检测速度的同时有效提升了模型对绝缘子的检测精度。
The insulator is an important part of high voltage transmission line,which is of great significance to maintain the stability of transmission line and ensure the normal operation of transmission network.Once the insulator breaks down,it will cause serious transmission failure and economic loss.For this reason,this paper proposes an improved YOLOv3 insulator detection method for high voltage transmission and transformation lines.Firstly,this method uses VoVNet with stronger feature extraction ability as the backbone network,and then proposes a new feature enhancement module,which can effectively enhance the semantic information and receptive field of the shallow feature map.The comparative experiments show that the proposed method mAP reaches 98.01%,which is 6.07%higher than the original YOLOv3 algorithm,thus not only ensuring the detection speed,but also effectively improving the detection accuracy of the model for insulators.
出处
《科技创新与应用》
2021年第34期79-82,86,共5页
Technology Innovation and Application
关键词
输电线路
绝缘子
深度学习
目标检测
transmission lines
insulator
deep learning
target detection