AI+ Digital Twin Road
2023.08.02
Project Name | AI+ Digital Twin Road |
Location | / |
Owner/Implementation Units | Implementation Units: Beijing Sinoits Tech Co., Ltd. |
Keywords | Intelligent transportation; Holographic visual perception; Traffic digital twin |
Completion Time | 2021 |
Description | Based on holographic visual traffic perception sensor technology and data, the AI+ digital twin road is a holographic traffic digital twin system integrating multi-sensor data covering GPS, microwave, radar, ground sense, CO/VI, and weather. Its advantage lies in that the system can not only collect traffic parameters, abnormal events, road operation situations, road environment, and other macro information but also perceive the comprehensive traffic information of vehicles (license plate, brand, car model, color, type, etc.), pedestrians, the position of non-motor vehicles, speed, and direction in real-time based on such technologies as visual perception and deep learning. Based on deep learning target recognition and full-time-space coordinate mapping technology, the system can automatically identify vehicle targets in surveillance images and realize the real-time perception, analysis, and trajectory collection and recording of the real size and geographic coordinates of vehicles, pedestrians, and objects through a three-layer coordinate mapping model algorithm of pixel level, two-dimensional space, and geodetic coordinates. The system can also automatically identify and track the license plates, brands, models, colors, and types of vehicles by combining image capture and intelligent identification technologies. The system adopts real-time and comprehensive traffic perception technology to record the time-space trajectory of traffic targets and facilities in a true, accurate and complete manner. It can provide not only massive, detailed and accurate historical data for the intelligent collaboration of real-time traffic but also important fundamental support for such applications as traffic simulation, sand table deduction, accident analysis and evidence collection of intelligent transportation systems. Through the digital twin road, the system constructs a digital foundation for intelligent transportation to provide real-time road information services for vehicles, promote the rapid development of intelligent roads, and fully realize digital management of roads. |
Key Data | Based on a new generation of artificial intelligence video recognition technology, the system can accurately identify as many as 7,000 styles of vehicles and the multiple features of more than 20 types of vehicles, covering almost 99% of vehicles on the road. Besides, it can construct 4,000-dimensional feature vectors for vehicles to realize the search for pictures by pictures. |
Stakeholders | |
Status Quo/Progress | The project has been completed. Relevant results of the project have been successfully applied in urban roads and expressways in Beijing, Guangxi, Jiangsu, Guangdong, Shandong and other provinces. |
Innovation Point | AI+Digital twin road combines deep learning high-definition video analysis technology and digital twin application big data platform technology. As the fundamental project for the realization of intelligent roads and transportation in the future, it can provide real-time, accurate and comprehensive vehicle-road perception, vehicle-vehicle perception, road-road perception, intelligent control, abnormal warning and other comprehensive traffic information. |
Related Links | |
Download attachment |
