Implementation of YOLO-v5 for a Real Time Social Distancing Detection

The world is in an uproar with the Covid-19 pandemic, which has had impact on society. Various efforts have been made by governments around to suppress spread of virus. One health protocols that appealed government social distancing or restrictions, namely limiting interactions between human beings as long 1-2 meters. But reality, there are still many people who ignore policies. application a detection system can be solution this problem. This aims detect violating form and then issue voice warning keep their distance from others avoid virus using YOLO-v5 method latest version YOLO (You Only Look Once). . Processing speed increased drastically fastest reaching 140 Frames Per Second (FPS) small size, even having size 90% compared previous version. accuracy reaches 83.28% 90.8%. From results percentage analysis carried out, it concluded created function well for detection, but difficult if humans too far camera.