Crowd Density Estimation and Real-time Monitoring using Digital Twins in Transportation Hubs

Digital twins transcend static models, offering dynamic virtual replicas of physical entities, thereby revolutionizing data-driven decision-making across industries. This paper proposes a novel digital twin system specifically designed for real-time crowd management, leveraging object detection with YOLOv8s architecture and 3D plotting using a three.js library. Our proposed system can be used as an edge analytics tool to minimize latency and enhance real-time decision-making, crucial for applications like autonomous vehicles, industrial automation, and large crowd gatherings. The proposed system was implemented and tested on multiple pre-recorded video feeds and it was evaluated based on time to process frames. For the proposed system an accuracy of 84.21% was achieved by comparing the actual number of people present in the given frame vs the number of people detected. This research demonstrates that digital twin models provide a robust, proactive approach to optimizing operations, enhancing safety, and increasing visitor satisfaction in crowded spaces.