Intelligent Visual-IoT-Enabled Real-Time 3D Visualization for Autonomous Crowd Management

Real-time crowd management systems play an important role in urban planning, disaster warning, and emergency evacuation. For example, through closed-circuit television (CCTV) cameras and artificial intelligence (AI) algorithms, crowd counting and abnormal event recognition can be realized. However, most of these applications are limited within the 2D coordinate system of each single camera, lacking a unified understanding of the 3D world as a whole. In addition, overly simple functionalities, such as framing certain areas or displaying text, limit the system intelligence and user experience. This article proposes a solution of realtime 3D visualization of outdoor scenes enabled by AI and the Visual Internet of Things (V-IoT). First, fixed and airborne cameras, infrared cameras, LiDAR, gas sensors, and so on are deployed to collect multi-modal data and send them to the clouds. Then AI algorithms are executed in the clouds to calculate the location of people, vehicles, and other moving objects, and to detect fights, stampedes, fire, or other abnormal events. Finally, the clouds send the AI algorithm results to the visualization system in terminal devices. The system displays the scene in real time through 3D animation and charts, and provides user interaction interfaces and simulation functionalities in a gamification environment. Experiments show that the proposed solution realistically visualizes the real world to sup-nort the user's decision.