The onset of the Covid19 Pandemic brought along a need for strict sanitation habits and social distancing. I created a plug and play solution which can detect crowding and count people. all the while actively calculating the distance between each person.

I used Yolov3 as the object detection framework for this project and was able to obtain a bird's eye view using perspective transform. This warps the image in such a way that, the warped image appears to be a bird's eye view of the place, and this enables us to calculate distances accurately.

The minimum distance is set to 1.8m, and this program creates a red bounding box around any two or more people who are at a distance less than the aforementioned distance.