Research on External Defect Detection Method of Transmission Line Insulators Based on Improved YOLOv7
Insulators are one of the important components of power lines.Due to long-term exposure to the outside environment,such as wind,sun,rain,and snow,it is inevitable to have defects such as damage and serious pollution.The operation of defective insulators will probably harm the normal operation of the power transmission system,so the detection of defective insulators plays an important role in the safety and stability of the power supply system.Aiming at insulator surface defects,an insulator defect detection method based on self-calibrating convolution YOLOv7 is proposed.Based on the YOLOv7 network model,the method replaces part of the conventional convolution by self-calibrating convolution,enhances the feature extraction capability of the model,expands the receptive field,and replaces the CIOU loss function of the original model with SIOU loss function to correct the insufficient box selection capability of the detection box.After training and detection on the existing data set,the final accuracy is improved by 5%compared with the original YOLOv7 model,which reduces the false detection rate of insulator defects of transmission lines and improves the efficiency of line inspection.
NETL
NSTL
万方数据
