Research on Insulator Defect Detection Method Based on Image Processing and Watershed Algorithm

Disc-shaped suspension insulators are widely used in substations and transmission and distribution lines of various voltage levels. Its main function is to achieve insulation and wire fixation in high-voltage overhead transmission and distribution lines. During operation, in a complicated environment such as pollution, temperature and humidity, insulator damage and chip-dropping failures often occur. The exploration of non-contact insulator defect detection methods has important theoretical significance and practical value. This paper takes the visible light image of suspended insulators as the research object. In this paper, the image region is extracted according to the shape and texture of the insulator, and according to the difference of computer vision parameters of various defects, an improved watershed algorithm is selected to realize the segmentation of the insulator image. Aiming at the possible defects such as breakage and chipping, this paper proposes the insulator defect detection criterion, so as to realize the automatic detection of the defect and the classification of the defect type. The research algorithm considers environmental factors and the efficiency and accuracy of algorithm feature extraction.