Computer vision is an important development direction of artificial intelligence, which can greatly improve the way people interact with the world. The basic elements of computer vision include data, algorithm models and computing power. The core technologies involve image classification, object detection, target tracking, semantic segm entation and instance segmentation. The computer visual algorithm technology layer has high barriers, the hardware layer is densely populated, the giants are numerous, and the application layer market is relatively scattered. The technology level of China's middle and upstream of the industry needs to be improved, white technology integration innovation, new algorithm development and cost reduction, GPU chip technology and industrialization are still Faced with severe challenges. In addition, front-end intelligence, front-end collaborative computing and software and hardware integration have become a clear development trend. Data and application scenarios will become the key points of enterprise layout. Deep learning and convolutional neural networks will promote continuous optimization and upgrade of computer vision. In the future, China should build a local computer vision industry cluster from the aspects of strengthening the coordinated development of the industrial chain, vigorously carrying out technical research, and supporting the application of future scenarios, and enhance the international competitiveness of the industry.