Track moving object in a given video.
- Numpy >= 1.24.3
- OpenCV >= 4.9.0 (latest as of 2024)
We have implemented Object Tracking Software which tracks moving object in a given video. We have implemented in two ways:
- From scratch using Lucas Kanade Tracking algorithm which uses six parameter affine model and recursive Gauss-Newton algorithm. Implemented the research paper Lucas-Kanade 20 Years On: by simon Baker (Microsoft Computer vision researcher).
- using OpenCV library.
Then, We have done the detailed comparative analysis of both implementation which can be found in the attached pdf.
- Put video in
video/
folder, or you can use provided sample videoslow_traffic.mp4
. - Run:
or
python3 lucas_kanade_scratch.py
python3 lucas_kanade_opencv.py
- Enter video filename along with path like:
Enter "0" to use your real time system camera for video.
Enter video filename along with path: video/slow_traffic.mp4