PF_video_tracking

Final project for the Applied Estimation course at KTH. We use a Particle Filter estimation algorithm to track the head of "The Annoying Bird" charter in a scene of "Kalle Ankas Jul".

Files in this Repositories

In this repository you will find the following files and folders:

• annoying_bird.mov: The video we are going to analyze.

• computer_vision.py: Contains the implementation of the computer vision functions used to extract the measurements.

• particle_filter.py: Contains the implementation of the functions related to the Particle Filter.

• visualizer.py: Contains the implementation of the functions that invokes the Particle Filter on the video.

• simulate.py: The main function that runs the algorithm with the selected parameters.

• test_cv.py: tester script used to debug computer_vision.py.

• test_pf.py: tester script used to debug particle_filter.py.

• test_vis.py: tester script used to debug visualizer.py.

• frames/: folder where frames can be saved.

How to run the Code

To run the project, open simulate.py and change the parameters at your pleasure. Then, run the same file with

python3 ./simulate.py

Pressing spacebar the video will stops for 10 seconds. In this situations, you can press s key to take a screenshot, that will be saved in frames/.

Maintainers

Alex Pegoraro

Theresa Johansson