Accurate Ultraviolet Image Detection of Electrical Equipment Based on Gaussian Color Mapping Image Segmentation Algorithm

Ultraviolet (UV) imaging technology has been used in corona discharge detection by characterizing the discharge degree through the detected number of photons in the UV imager or facular area of the UV image. This paper is to expand the UV imaging method in the electrical equipment’s fault detection and improve the accuracy of its detection. First, an image segmentation algorithm based on Gaussian function color mapping is proposed for the segmentation of colored facular areas in the UV image, and its effectiveness is proved. Second, the testing results for a high-voltage insulator’s discharge are used to fit the distance attenuation function of the facular area. The relationship between the voltage and the attenuation function coefficient is obtained by analyzing the experimental results. Finally, by analyzing actual inspection data of insulators in a substation, it is found that different defect types of insulators have different characteristics of changes in the facular area during the discharge. Based on the characteristics, accurate direct detection by using UV imaging can be realized for the defect types of discharging insulators.