Skip to content

AissamDjahnine/Mouvement_estimation

Repository files navigation

Optical Flow Estimation with Horn-Schunck Method and Multi-Resolution

Experience Horn and Schunck's groundbreaking work, the first variational method for optical flow estimation. Our project implements their method and introduces a multi-resolution strategy to handle larger displacements. By assuming constant pixel intensities over time, we derive the optical flow equation. While effective for small motions, this method struggles with larger displacements due to image gradient issues. Our approach involves a smoothness constraint and a multi-scale pyramid of downsampled images for robust optical flow estimation.

Usage :

Inputs : Image 1 & Image 2 ( un mouvement elementaire )

Image 1

alt text

Image 2

alt text

Resultats (via quiver pythonb) :

alt text

About

Mouvement estimation ( Horn & Schunk Method ) C Programming

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published