detecting crossing line and presence on virtual fences using object tracking models.
In this project, I implemented the detection of the crossing a virtual line and presence on the virtual fences using object detection and object tracking models. The coordinates of the line and virtual fences were given in the policy.json file.
The python libraries that need to be installed is shown in requirement.txt.
The result for detecting crossing line is shown as below:
output_lines.online-video-cutter.com.mp4
As you can see, this code, each person is detected with a bounding box and assigns a unique ID for each person in the video (tracking part) and after crossing the line, the line color is changed to red and remains until 3 seconds.
For reaching this result, Run the line_count.py. Enter your custom video path in the code and also the coordinates of the lines.
The result for presence on virtual fences is shown as below:
output_fences.online-video-cutter.com.mp4
Same as preveious, If a person is present on the virtual fence, the color of the fence will change to red and the number of people present on the fence will also be displayed. And if there is not any person in the fence, the color will be green and the number will be zero.
For reaching this result, Run the fences.py. Enter your custom video path in the code and also the coordinates of the fences. Another feature of this code is that you can move the virtual fence area with a mouse click and move it to different locations in the video and have the number of person that are present in the fence.
If you have any question, please don't hesitate to ask me. Email: [email protected]
And if this repository was helpful, please star it.