A lightweight detection method for insulator self-explosion based on YOLOv5s
Insulators play a key role in power lines and are vital to ensure the safety and reliability of the power system.Due to their vulnerability to adverse weather conditions,insulators are required to carry out regular inspection.With the advancement of artificial intelligence,the use of unmanned aerial vehicles(drones)for dynamic inspections has become increasingly popular.It can use aerial photography to obtain insulator images in a more convenient way.However,there exist some challenges for use algorithms to recognize fault insulator images.For instance,small targets are difficult to be recognized and easy to be cover.Moreover,dynamic insulator inspections needs to achieve a balance between accuracy and real-time capabilities.Therefore,this paper proposes a detection method of self-explosion of lightweight insulators based on YOLOv5,which can increase realize rapid and accurate dynamic detection of insulator self-explosion by improving the loss function,optimizing the detection model,and enhancing inspection accuracy.
NETL
NSTL
万方数据
