Electric power defect image recognition method based on neural network model

The invention relates to the technical field of electric power system defect detection, and discloses an electric power defect image recognition method based on a neural network model, which comprises the following steps of: 1, firstly, acquiring images of electric power equipment, a transmission line and an insulator on site, and inputting the acquired images into terminal processing equipment; and marking different power equipment, transmission lines and insulator images by using terminal processing equipment. According to the method, the terminal processing equipment is used for marking different power equipment, transmission lines and insulator images, positions and time of the different power equipment, transmission lines and insulator images can be marked, and after the power equipment, transmission lines and insulator images are processed, the positions and time of the power equipment, transmission lines and insulator images can be marked. The positions and the time of the images of the power equipment