Study On Detection Method For Internal And External Defects Of Overhead Line Insulators

With the rapid development of economy,the demand for power resources in China is growing,and the scale of transmission lines is also expanding.Insulators,as an important part of the transmission line,have been in a harsh environment for a long time with a high failure rate.If the faulty insulators in the transmission line cannot be checked in time,the normal operation of the transmission line will be affected.The traditional manual inspection of transmission lines has problems such as low efficiency and poor safety.With the rapid development of UAV and computer vision technology,the intelligent defect detection algorithm of insulators has become a research hotspot.In this paper,various types of insulators are specifically analyzed.Three common defects,namely,porcelain insulator shed falling off,glass insulator self explosion,and composite insulator abnormal heating,are studied.Two detection algorithms,MDD-YOLOv3 and Infra-Detect,are designed,and an insulator defect detection system is built,as follows:(1)A lightweight defect detection method MDD-YOLOv3 is proposed to solve the problems of ceramic insulators and glass insulators,such as close arrangement,complex back ground and small defect area.First,a new backbone network,DDarknet53 is established to reduce the network parameters,and the Dense-SPP module is designed to improve the network’s ability to express the characteristics of defects.Build a four scale prediction layer to improve the small target detection performance of the network.The detection accuracy of the improved network MDD-yolov3 for insulators reaches 96.1%,and the detection speed reaches 36 frames · s-1.Compared with yolov3,the detection accuracy and speed are improved by 4.0% and28.6% respectively.(2)In view of the internal defect of composite insulator abnormal heating,an insulator thermal defect detection model Infra-Detect is proposed by using infrared images.First,the improved image segmentation algorithm G-Grabcut is used to extract insulators from the infrared image.Then,according to the pixel value and temperature information of the colorimetric bar,the functional relationship between the temperature of the infrared image and the pixel is fitted,so as to read the temperature of the insulator.Finally,the insulator thermal defect grade is diagnosed by using the relative temperature difference method according to the insulator thermal defect judgment standard.The experiment proves that Infra-Detect can accurately segment the infrared image of insulator with low brightness and complex background,and accurately diagnose the thermal defect level of insulator.(3)The insulator defect detection system is designed,including UAV image acquisition system and detection system software.The visible light and infrared images of insulators are acquired through the UAV image acquisition system,and the MDD-YOLOv3 and Infra-Detect models in the system software are respectively used to identify the surface and internal defects of insulators.Figure [41],Table [10],Reference [82]...