learnopencv/Optical-Flow-in-OpenCV

Optical Flow in OpenCV(C++/Python)

This repository contains the code for Optical Flow in OpenCV (C++/Python) blogpost.

Python

Installation

Before you start the demo of Optical Flow calculation, you need to create a virtual environment in your working directory and install the required libraries:

virtualenv -p python3.7 venv
source venv/bin/activate
pip install -r reqirements.txt

Sparse Optical Flow

There is a demo lucas_kanade.py script of Lucas-Kanade algorithm which can be run with this command:

python3 demo.py --algorithm lucaskanade --video_path videos/car.mp4

Dense Optical Flow

The wrapper of Dense Optical Flow algorithms dense_optical_flow.py can run a couple of OpenCV's algorithm implementations:

• To start the Dense Lucas-Kanade algorithm: python3 demo.py --algorithm lucaskanade_dense --video_path videos/people.mp4
• To start the Farneback algorithm: python3 demo.py --algorithm farneback --video_path videos/people.mp4
• To start the RLOF algorithm: python3 demo.py --algorithm rlof --video_path videos/people.mp4

C++

Installation

Before you start the demo of Optical Flow calculation, you need to build the project:

cd algorithms
cmake .
make

Sparse Optical Flow

There is a demo lucas_kanade.cpp script of Lucas-Kanade algorithm which can be run with this command:

./OpticalFlow ../videos/car.mp4 lucaskanade

Dense Optical Flow

The wrapper of Dense Optical Flow algorithms dense_optical_flow.py can run a couple of OpenCV's algorithm implementations:

• To start the Dense Lucas-Kanade algorithm: ./OpticalFlow ../videos/car.mp4 lucaskanade_dense
• To start the Farneback algorithm: ./OpticalFlow ../videos/car.mp4 farneback
• To start the RLOF algorithm: ./OpticalFlow ../videos/car.mp4 rlof

AI Courses by OpenCV

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.