Insulator UV Image Fault Detection Based on Deep Learning

In this paper, ultraviolet images of porcelain insulators were taken as the research object, a sample library of ultraviolet images of insulators with different insulation states was established, and a convolutional neural network model was built. The model was operated under different learning rates using AlexNet structure, and the training process was analyzed visually. The results showed that the learning rate had a great influence on the performance of the network, and the accuracy of the test set reached the optimal value under a certain learning rate. Through supervised learning of samples, Convolutional Neural Network realized autonomous feature extraction and high-precision fault evaluation for UV imaging of porcelain insulators.