Real-Time Crowd-Counting and Management in Sacred places using Computer Vision & ESP32 cameras
Mohammed Mahmood Ali, Mohammad Sanaullah Qaseem, Mohd Zeeshan, Abrar Ahmed Quraishi, Ateeq Ur Rahman · 2023
Flexible solution for real-time crowd counting and control in holy places that provides flexibility of an ESP32 camera and the accuracy of YOLOv8 object detection framework. It is designed to ensure crowd safety and prevent potential stampedes in dynamically changing religious environment. ESP32 cameras strategically placed throughout the sanctuary, recording live video feeds, allowing easy relocation based on changing audience dynamics on different days or times of the week. These video streams are forwarded to a central server for efficient processing. The server preprocesses and uses the YOLOv8 model to accurately count the number of people in each area. The system includes actuators such as automatic motorized door controls and information LCD screens to direct visitors to less crowded areas. The proposed solution addresses the unique challenges facing holy places by offering a comprehensive and customizable approach to crowd counting and management, thereby improving the safety and sanctity of religious settings. The precision obtained when compared to state-of-the-art systems was 68.9%