AN AI BASED SOCIAL DISTANCE MONITORING SYSTEM

Abstract

The virus that causes Covid19 is identified as SARS-CoV-2. The corona virus’s devastating spread has brought forth a global catastrophe. Social isolation is thought to be a defense mechanism against the pandemic virus’s rapid spread. By avoiding direct social contact with other people, the danger of the virus spreading can be reduced. In order to develop a deep learning platform for social distancing utilizing an aerial perspective, this work’s goal is to achieve this. To detect people in video footage, the framework employs the YOLO object detection approach. The recognition and tracking of individuals in both indoor and outdoor environments is done using a deep learning detection approach. The users of the discovered bounding box information are identified using the detection model. The distance between two individuals from the centre of the observed bounding box is calculated using the Euclidean distance. Here utilised a pixel-physical distance calculation and a threshold to calculate the prevalence of social distance violations between individuals. To determine whether the distance value exceeds the minimal social distance criterion, violation thresholds are developed. Additionally, a tracking algorithm is employed to identify people in the video clip in order to follow anyone who violates or crosses the social threshold. The suggested method may be used for a low-cost embedded device with a fixed camera. The suggested method may be used to watch individuals from various cameras in a centralized surveillance system using a distributed CCTV system. This method is appropriate for establishing a surveillance system in smart cities to find individuals, categorize them, and assess social distance.