Image Recognition based on Convolution Neural Network

This paper introduces the development of image recognition technology, summarizes several classic traditional algorithms, and then introduces the use of convolution neural network for image recognition, and summarizes the application and development of CNN in various fields in recent years, and proposes to recognize images by deep learning in complex or high-risk environment similar to substation Don't take the place of manpower to detect the working state of instruments and equipment, which has a development prospect. In the end, it puts forward some suggestions on the direction of image recognition using deep learning. In the future, we need to improve and optimize the model, intelligent algorithm and data set.