ABSTRACT
The world has shrunk due to the COVID-19. This virus took many lives of people. Hence the government proposed a lockdown and closed the educational institutions to decrease the death rate. The cases are controlled to some extent, but still the spread of the virus is not yet controlled. Though lockdown has been released and the educational institutions are being slowly opened with some conditions like wearing masks, maintaining social distance in the public areas and avoiding contact with the infected persons. But the people are not maintaining social distance in the colleges and schools, which can cause risk. Hence it is mandatory to build a model which identifies the people who cross the safe limit of distance. This work focuses on determining the social distance in the educational institutions mainly in the colleges. Generally, most of the colleges have Closed Circuit Television (CCTV) cameras fixed on the campus, in order to determine the social distance among the people, the videos are taken from those CCTV cameras. The people are detected from the video clips using the You Only Look Once Version 3 (YOLO V3) algorithm which is trained on the Common Objects in Context (COCO) dataset. And the safety distance among the people is found using Euclidean distance. The minimum threshold value is fixed to the safe distance suggested by World Health Organization (WHO). Moreover, a warning is sent to the corresponding coordinators through Short Message Service (SMS) when the people are not in the safer social distance. © 2022, Ismail Saritas. All rights reserved.