Identification Method for Risk of Electric Water Flushing in Non-Dense Areas of Ice-Covered Insulators

The operation of live water flushing itself poses safety risks, especially under ice-covered conditions, where icicles and other substances may exist on the surface of insulators, increasing the complexity and danger of flushing. To this end, a risk identification method for electric water flushing in non-dense areas of ice-covered insulators is proposed. The neighborhood pixel gray value average method is adopted for the smoothing and denoising processing of the non-dense area. The maximum betweenness variance method is selected as the threshold segmentation method. The interference area is eliminated through area comparison, and the area to be washed is segmented from the non-dense area. Multiple factors such as water quality, washing operation method, equipment type and structure, environment and meteorological conditions are comprehensively considered. Comprehensively analyze the flashover risk in live water flushing. By calculating the intersection and union ratio between the detection bounding box and the existing target bounding box or the Euclidean distance of the center coordinates, it is determined whether the detection bounding box contains new targets. Wavelet edge detection is adopted to extract the edges of ice-covered insulators, and the multi-scale characteristics of wavelets are utilized to achieve noise suppression and reliable edge identification. The experimental results show that the success rate of the average pixel error (20px) of the tracker in the proposed method is 84.8%, the error rate of the upper insulating sheet is 1.5%, and the error rate of the lower insulating sheet is 1.7%, all of which remain at a relatively low level and can accurately locate the insulator.