Research on Insulator Defect Detection and Recognition in Overhead Transmission Line Based on Convolutional Neural Network

Abstract

In the overhead transmission line, the insulator plays a very important role in the operation of the power system. Because the insulator is exposed to the field for a long time, the insulator is often damaged, iced and hung with foreign matters. This paper studies the defects of the insulator in the overhead transmission line. In overhead transmission lines, human eyes cannot directly observe the defects of insulators. Insulators in overhead transmission lines are often photographed with the help of unmanned drone, helicopter and other auxiliary equipment, and then identified by different methods. In order to improve the accuracy of insulator defect identification in overhead transmission lines, this paper uses hierarchical detection method to detect and identify insulators. Firstly, fast r-cnn algorithm is used to locate insulators and cut insulators, so as to increase the proportion of insulators in the image. Secondly, image enhancement method is used to expand the data set, finally, yolo-v4 algorithm is used to detect the defects of insulators. In order to improve the speed of insulator detection in overhead transmission lines, resnet50 is used as fast r-cnn feature extraction network in the first level detection; In order to improve the identification accuracy of insulator defects in overhead transmission lines, mobilenetv2 is used as the feature extraction network of yolo-v4 in the second level detection. Experimental results show that this method can effectively and accurately identify and detect insulator defects in overhead transmission lines.