Ultraviolet Image Segmentation and Characteristic Quantities of Corona Discharge Based on C-V Model

When corona discharge happens in high voltage equipment due to insulation defect, it generates ultraviolet (UV) light. So UV imaging technology can be employed to detect the position and intensity of corona discharge and to assess the insulation defects. UV imager provides UV photon number to indicate the discharge intensity. However, the photon number becomes inaccurate due to environmental disturbance. In this paper, the C-V level sets model is adopted to segment UV image to extract discharge characteristic quantities for the purpose of more accurate description of discharge characteristics in high-voltage equipment. Firstly, C-V level sets model is introduced to segment UV image to extract discharge facula region and coordinates of the region boundary points. Then, according to the extracted region and coordinates to determine such characteristic parameters as facular area, perimeter, major axis, and minor axis for evaluating discharge state, and the correction method of interference factors is given. Practical examples of processing UV discharge images of a substation insulator demonstrate that this C-V model based method can accurately segment the discharge area without filtering preprocess and with high automation, and it is adaptable to UV images with various backgrounds.