Social and Safety Monitoring for Pandemic with YOLO

Abstract

During the pandemic of COVID-19, people have been suggested to keep social distance from the others. It is also beneficial to pay attention to the individuals with motion irregularities. In this research, we propose a visual anomaly analysis system based on deep learning with the aim to identify individuals with various types of anomaly which are specifically important during the pandemic. Based on the proposed monitoring system it would be easier to keep tracking the environment changes, and it would also be beneficial to the safety guard to reallocate resources accordingly to relieve the threat of anomaly. Types of the anomaly are very sensitive during the coronavirus pandemic. In the study, two types of anomaly detections are concerned. The first is monitoring the abnormally in the case of falling down in an open public area, and the second is measuring the social distance of people in the area to keep warning the individuals under an insufficient distance. By the implementation of YOLO, the related anomaly can be identified accurately in a wide range of open area. The reliable results make promisingly the use of a vision sensor as a ranger to detect anomaly in time in the open area. Through the implemented system to monitor the environment, the safety monitoring would be easier to manage the anomaly around a neighborhood which may help to avoid the spread of the virus.