Ultraviolet (UV) imaging is an effective visual discharge detection method. In this paper, the ceramic insulator pollution discharge character was researched with the UV imager in the artificial climate chamber. According to the discharge UV image, the discharge type is divided into two typical forms, one is corona discharge, and another is the partial arc discharge. Using digital image processing algorithm, the discharge UV image region was segmented from the background image, the partial arc discharge number in certain time was counted, and the correlation of these parameters with the humidity were researched. Fuzzy logic inference was applied to realize the pollution grade evaluation. Application showed that fuzzy logic inference method can effectively to assess the pollution grade.