https://github.com/alejo1630/video_tracking
This Python Notebook allows you to perform a video tracking of an object and obtain its kinematic information such as displacement, speed and acceleration.
https://github.com/alejo1630/video_tracking
computer-vision jupyter-notebook kinematics python video-tracking
Last synced: 5 months ago
JSON representation
This Python Notebook allows you to perform a video tracking of an object and obtain its kinematic information such as displacement, speed and acceleration.
- Host: GitHub
- URL: https://github.com/alejo1630/video_tracking
- Owner: alejo1630
- Created: 2022-12-05T17:21:26.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-12-16T21:57:04.000Z (over 3 years ago)
- Last Synced: 2025-07-19T23:58:11.521Z (12 months ago)
- Topics: computer-vision, jupyter-notebook, kinematics, python, video-tracking
- Language: Jupyter Notebook
- Homepage:
- Size: 2.05 MB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Video Tracking Kinematics
This Python Notebook allows you to perform a video tracking of an object and obtain its kinematic information such as displacement, speed and acceleration.
## 🔰 How does it work?
- The code uses [OpenCV library](https://opencv.org/)
- User must upload the video.
- Some information is obtained from the video (FPS, duration, resolution)
- The first frame of the video is extracted in order to create a reference line to convert the pixel data into measurement units (i.e. cm).
- The reference line is created with a mouse_callback event where the pixel coordinates of the start and end point of the reference line are used for the unit conversion. For a correct measure, the video must have a reference guide such as a flexometer or ruler. The user must be knonw the real distance (cm, inches, etc) of the reference line. The current code is based on X direction but it could be modify to Y direction.
- In the next step, the user must create a bounding box around the object to be tracked.
- After creating the bounding box the tracking process will start. Whenever the program succeeds in tracking, a green message "Tracking" will appear, otherwise a red message "Lost" will be displayed.
- When the video ends, all the kinematics of the object's movement are calculated.
- Displacement
- Speed
- Acceleration
- Based on the quality and FPS of the video, the data could have some noise. This is treated by two methods
- [Savitzky–Golay filter](https://en.wikipedia.org/wiki/Savitzky%E2%80%93Golay_filter)
- [Spline Interpolation](https://en.wikipedia.org/wiki/Spline_interpolation)
- The results obtained are compared with the video data processed in the [Tracker software](https://physlets.org/tracker/)
## 🔶 What is next?
- Improve the accuracy of the kinematics data.
- Perform the same analysis for live videos.