The Vision Information Vehicle Refinement Recognition System effectively combines image intelligence analysis and laser detection 3D modeling technology; The operation uses an HDFS distributed file system based on Infiniband network.Special processing of laser acquisition data improves the accuracy of laser recognition and modeling of different color vehicles. Based on the vehicle positioning, the deep analysis technology is used to intelligently analyze the collected video data. The system matches the vehicle data results analyzed by the image intelligent analysis technology with the vehicle model data established by the laser three-dimensional modeling technology. Real-time output of the matched vehicle refined recognition results to form an intelligent transportation vehicle data sharing library. Vehicle refined identification data can further analyze the intelligent transportation business,Research and early warning, realize the refined management of intelligent transportation, and promote the development of intelligent transportation industry. The Pedestrian Attribute Recognition Technology of Vision Information uses deep learning techniques based on pedestrian detection and pedestrian sub-components and attachment positioning.Can achieve refined recognition of pedestrians. Vision takes the lead in applying deep learning technology to the field of intelligent video analytics. In the field of image understanding application, it has made a breakthrough change, realizing the application of artificial intelligence in the field of video processing and analysis. Video Concentration Abstract Technology and Video Retrieval Technology independently developed by Lai King, It can automatically perform daily diagnosis on the front-end image acquisition device, and perform intelligent summary retrieval on massive video, which greatly reduces the workload of personnel.