Electric power fittings are the critical components of power transmission and transformation structures. Under the action of environmental factors, such as temperature and sandstorm, structural defects can lead to corona discharge, which can ultimately affect the safe operation of power transmission and transformation lines. Based on the working platform of a corona test, the feature extraction and evaluation of discharge spots were carried out in this study using image fusion and image processing technology. The results prove that the relationship between the spot area and discharge capacity is linear. The discharge stages of electric power fittings develop from the corona stage to a small electric arc stage, and finally, to a strong spark discharge stage. When the observation distance is 10 m, the gain is 70%, and the discharge spot areas of three stages are approximately 1000, 5000, and 10000 pixels. Based on the characteristics of the ultraviolet image, the state of electric power fittings can be divided into four states: no defect, general defect, serious defect, and emergency defect. The study results provide technical support for the safety evaluation of electric power fittings.