SOCIAL DISTANCING DETECTION WITH DEEP LEARNING MODEL

Computer Engineering

The ongoing Covid-19 pandemic is a global disaster that disrupted the normal life of the people and caused more than 2 million deaths worldwide. It is important to monitor the social distance and wear masks at public places and take actions accordingly. This tool was developed to alert humans to maintain a safe distance with each other by video feed. Deep learning gained more attention in object detection which is used for human detection purposes. The video frame from the camera is used as input, and the open-source object detection YOLOv3 object detection is used to identify humans in video sequences. Then, the video frame is transferred into Bird-Eye view for distance measurement from the 2D plane. The distance between people can be estimated and any non-compliant pair of people in the video will be indicated with a red bounding box. If the people maintaining the safer distance, then it will show in green bounding box. By calculating the distance among individuals who are identified in the frame, the number of people who are at violating social distance is classified and indicated. The method will be able to determine the distancing measures between multiple people in indoor

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.