A Novel Approach for a Vision-Based Physical Distancing Monitoring System for COVID-19 Using Deep Learning Techniques

Abstract

AbstractNovel coronavirus (COVID-19) is one of the most contagious diseases in the era which was declared as a pandemic by World Health Organization (WHO). A majority of people do not have any symptoms and are the main reason for transmission of virus also known as “Super Spreaders.” The main ways of the spread of this disease are when an infected person coughs, sneezes, or talks, droplets from their mouth or nose are launched into the air and the droplets are inhaled into the lungs of people nearby. To tackle this, physical distancing must be followed. Physical distancing means keeping a safe space between yourself and other people. Physical distancing needs to be followed with all other preventive measures. There is a need for a system that monitors the physical distancing norms through which if a violation is detected, then based on the number of violations to the number of people, the system should alert and helps in imposing physical distancing measures. In our proposed system, we try to overcome the above-mentioned disadvantages by creating a physical distance monitoring system in real-time using a stationary camera through which video is captured using OpenCV and processed using YOLO v3 algorithm, through which we will alert the violations of the physical distancing and calculate the density and assess the risk by using statistical calculations that reports the location to the concerned authorities and make them follow physical distancing.KeywordsCoronavirusPhysical distancing monitoringOpenCVReal-time object detectionYOLO version 3Euclidean distance