Defect detection method of transmission line based on improved YOLOv5
Currently,power grid companies are gradually adopting unmanned aerial vehicles(UAVs)for intelligent inspections of electrical equipment,with a focus on detecting insulator defects,missing R cotter,and nesting.This paper proposes an improved YOLOv5 algorithm to enable defect detection in transmission lines.Firstly,the ultra-high-resolution images captured by the UAV are processed into smaller patches,and the dataset is analyzed using the K-means algorithm to obtain the optimal anchor box size.Secondly,BiFPN(bidirectional feature pyramid network)is adopted to replace the original feature pyramid network of YOLOv5 to realize more free information exchange and feature fusion.Finally,the SimAMattention module is added to solve the influence of complex backgrounds on defect identification accuracy.The experimental results show that the detection effect of the improved YOLOv5 algorithm is significantly improved,so it has high application value.
NETL
NSTL
万方数据
维普
