Low/Zero Fault Diagnosis of Porcelain Insulator Based on HSV Color Space Image Processing Technology

Abstract

In this paper, the typical disk suspension porcelain insulators in transmission and distribution lines were taken as the research object, the infrared image differences between low/zero insulators and normal insulators were analyzed, and the low/zero fault diagnosis of porcelain insulator was carried out by HSV color space image recognition technology constructively. Firstly, the prediction model of porcelain insulator deterioration was constructed by using YOLOv3 algorithm, and the smallest rectangular area of insulator was extracted from the infrared image. Then the insulator was segmented from the rectangular region by using K-means algorithm to avoid background interference. On this basis, the normalized values of HSV color space features, including the maximum and minimum values of H component, the average value of S component and the maximum value of V component, were drawn according to the long axis direction by using the rotating rectangle to correct the insulator image. Combined with the correlation between HSV components and temperature rise, the detection criterion of low/zero fault of porcelain insulator is deduced, the operation state evaluation of insulator is accomplished.