Research On Insulator Defect Detection Algorithm Based On Improved YOLOv5n

As my country’s electricity consumption continues to grow,high-voltage transmission lines continue to expand.In transmission lines,insulators are one of the important components,mainly used for electrical insulation and mechanical fixation.Because insulators are exposed to harsh environments for a long time,insulators on transmission lines are prone to failures such as burns and damages.These failures will lead to a decrease in their insulation performance,which will seriously affect the safe and stable operation of the power grid.In recent years,with the rapid development of artificial intelligence technology,insulator defect detection methods based on deep learning have been gradually applied to aerial insulator images,which not only improves detection efficiency,but also saves manpower.However,aerial insulator pictures are affected by some factors,such as shooting angle,shooting distance,and the complex environment in which the insulator is located,which makes it difficult for the insulator detection method to achieve the expected results.In view of this,in order to improve the detection accuracy of defective insulators,this paper adopts the improved YOLOv5n network,proposes an insulator defect detection algorithm,and designs an insulator defect detection system.The main work is as follows:(1)Aiming at the problem that the background of aerial insulators is complex and the defects are not obvious,this paper proposes an insulator and its defect detection algorithm based on the attention mechanism.This method introduces a new attention module MECA into the feature extraction network of YOLOv5n,so that the model can not only focus on the area where the insulator is located,but also perform multi-scale information fusion to avoid losing shallow information,thereby improving the accuracy of small targets.(2)In order to make the insulator defect detection model more lightweight and improve the detection speed,this paper proposes a lightweight detection algorithm for insulators and their defects.First,the backbone network is replaced by the lightweight model Ghost Net to reduce the number of parameters of the model.Next,considering that the selection of the activation function has a great influence on the performance of the network,this paper compares the two activation functions with better performance,and selects the activation function that is most suitable for the insulator defect detection model.Finally,aiming at the inaccurate box regression problem,a new loss function PEIOU is proposed as the box regression loss function to improve the overlap between the predicted frame and the real frame,and accelerate the convergence speed of the model.(3)Design a set of insulator defect detection system,which combines the insulator defect detection algorithm with the interface designed by PyQt5.It can detect pictures and videos,and output the detection results,making it easier for staff to access defective insulators.In short,this paper comprehensively improves the YOLOv5n model,and conducts recognition tests on the constructed aerial insulator dataset and public insulator dataset.The experimental results show that the improved model improves the recognition accuracy of insulator defect detection,can meet the requirements of insulator and its defect detection,and lays the foundation for subsequent engineering applications.