Social separation has proven to be a very effective way of slowing the spread of the disease in the fight against coronaviruses. Although millions of people remain at home to help flatten the curve, many consumers in the manufacturing and pharmaceutical industry also have to go to work on a regular basis to ensure they meet our basic needs.
Tool for monitoring social distance from CCTV, videos using Python, Deep Learning, Computer Vision. This tool can automatically estimate the interpersonal distance between uncalibrated RGB cameras. It can be used in public places and workplaces.
The three key steps of the approach:
- Detect people in the yolov3 neural network in the frame.
- The distance between all instances of people in the frame can be calculated.
- Classifies distances for social distancing as ‘Alert’ or ‘Ok’.
[adinserter block=”2″]
Requirement for Social Distance Dectector
- Download yolov3.weights for COCO dataset click here
- Numpy
- OpenCV
- OpenCV_Contrib
- Math
Model Installation
- To deploy on images use SDD_Image.py
- To deploy on videos use SDD_Video.py
- To deploy on live streaming use SDD_Camera.py
[adinserter block=”2″]
[su_button id=”download” url=”https://drive.google.com/file/d/199ZyUNJARLC08xOqP2p2JbXxsEohaCFs/view?usp=sharing” target=”blank” style=”flat” radius=”10″ icon=”icon: files-o”]Download Now[/su_button]
Leave a Comment