International Research Journal of Education and Technology

To prevent the spreading of COVID, the only way is social distancing. Nowadays, AI teams create social distancing tools using computer vision concepts. This project proposed a methodology to find social distance with the help of deep learning to evaluate the distance between people for mitigating the impact of the coronavirus pandemic. The detection tool was developed to notify people to keep a safe safety distance from each other through the evaluation of a video input feed. The video frame from the 'mp4' file was given as input, and an object detection pre-trained model based on the YOLOv3 algorithm was applied for pedestrian detection. Then, the video frame was converted into a top-down view to measure the distance from the 2D plane. The distance among people was estimated and any non-compliant pair of people in the display is indicated with a red frame and red line. The proposed method was validated on pre-recorded video of pedestrians walking on the street. The output result verified that the proposed method can determine social distancing measures between people groups in the video. The developed technique may be further developed as a detection tool in real-time application. The project is designed using Python 3.5 with Open CV python 4.2.

Keywords: Social distance monitoring, Covid 19, human object detection.