A script to track objects and estimate their speed in videos that move in a straight path perpendicular to the camera.
The script uses OpenCV to track the objects and exports the collected estimated speed data to a csv file.
An OpenCV 3.4+ installation is needed. Either build it from source (installation guide for Ubuntu) or install it with:
pip3 install opencv-contrib-python
- Run program:
python3 speed_tracker.py video_in csv_out frame_time [-l reference_length] [-r rotation]
-
Press r to rotate the video clockwise by 90 degrees. If the video is aligned press space.
-
Click and drag to draw the reference line with the specified length. By default the reference length is 2 meters. Press space to continue.
-
Click and drag to select the object you want to track. Press space to continue.
Now the tracking starts. The position of the object is cached every frame.
You can now control the analysis of the video as follows:
-
Pause / resume the video with space
- When the video is paused or over, press s to save cached data
-
Clear the cached data with r
-
Exit the video with esc