Real-Time Surveillance System for Detection of Social Distancing

ABSTRACT

As the corona virus can mutate and due to other scientific factor associated to it, experts believe that COVID-19 will remain with us for decades. Therefore, one has to keep social distancing measures. Accepting the pandemic situation, the paper presents a mechanism for detecting violations of social distancing using deep learning to estimate the distance between individuals to diminish the influence of COVID-19. The focus of this paper is to understand the effect of social distancing on the spread of COVID-19 by using YOLOv3 and Faster-RCNN and proposes IFRCNN (improved faster region convolution neural network). The proposed method IFRCNN is checked on a live streaming video of pedestrians walking on the street. This paper keeps the live updates of the recorded video along with social distancing violation records on a location, so how many people in a location are maintaining social distancing. Updates will be stored in a cloud-based storage system and any organization or firm can get live updates of that location in their digital devices.