Smart Crowd Monitoring System using Image Processing Technique During Covid-19

ABSTRACT

During the COVID-19 scenario, due to partial lockdown, people did not have permission to go out and buy items freely. Instead, they were given a specific time for purchasing goods. As a result, people were found in multitudes during these hours, without maintaining social distance. Managing this crowd to maintain social distance is a huge task for the government, and hence a system that will assist them in controlling the people is required. The You Only Look Once (YOLO) approach was used to detect the objects. Compared to other object detection methods, this technique has a lot of advantages. YOLO finds objects by applying convolutional networks to forecast bounding boxes and class probabilities for these boxes, and it does it much faster than the existing works. This paper develops a device using a Raspberry Pi-4 board that detects people who are in the frame of the camera, and if they are closer than the distance allocated in the device, an alarm will sound, informing them that they are breaking the rules, and the alert message will be sent to the nearby police station. In this way, the crowd can be managed in a pandemic situation. © 2022 IEEE.