Lucas-Kanade Optical Flow

This is a very slow, toy implementation of the Lucas-Kanade Optical Flow algorithm for self-learning purposes. Methods including 2D convolution, taking image gradients and the least squares approximation of flow velocities were done from scratch.

See example output below:

Usage

1. Clone the repository

git clone https://github.com/aswinvisva/optical_flow_algorithms.git

2. Make a build directory

mkdir build
cd build

3. Generate the MakeFile using CMake

cmake ../

4. Build the project

make

5. Run with a video of your choice

./OpticalFlowAlgorithms --video_dir