IDamage Detection and Identification of Insulators Based on Improved Yolov5
Insulators are applied in carrying voltage and mechanical pressure on conductors with different potentials.Due to the influences of electrical environment and power load fluctuations,insulators may be subjected to multiple electromechanical coupling stresses,which does not work properly and affects the lifespan of entire insulator networks.To address this issue,an insulation dam-age detection algorithm through object detection is proposed.The improved scheme is based on the Yolov5s model.Firstly,more small object detection layers were added to the original Yolov5s model,thereby improving the detection accuracy.In addition,an ad-ditional operational layer was introduced to extend the feature map,and the SE(Attention and Observation)attention module was used to make the network more focused on detecting objects.SIOU was also used to replace the loss function in the YOLOv5s.The experimental results show that the improved model has significant advantages in insulator damage detection compared to the traditional Yolov5s model.The improved model improves the mean accuracy(mAP),precision(P),and recall(R)by 2.5％,1.1％,and 0.8％,respectively.Compared with the original Yolov5s model and other models(such as Yolov5m,Yolov51,etc.),the improved model has stronger competitiveness in insulator defect detection and recognition,provides an effective solution for improving the accu-racy of insulator damage detection,accurately detects insulator damage and early necessary repair and maintenance measures,and ex-tends the lifespan of insulators and ensures the stable operation of the power system.
NETL
NSTL
万方数据
