Visual SLAM

Introduction

Code for implementation of Visual SLAM using data from an RBG Camera equipped on a Autonomous Vehicle using Eight Point Algorithm for the 3D Point Clouds data.

Overview of the Repository

In this repo, you'll find :

• Kitti: famous kitti dataset.
• hartley1997.pdf: paper describing Eight Point Algorithm for 3D Point Clouds data.
• motion2D2D.py: motion estimation from 2 frames in a monocular setup.
• draw.py: modified version of opencv DrawMatches to recover randomly generated colors.

Getting Started

1. Clone repo: git clone https://github.com/HusseinLezzaik/Visual-SLAM.git
2. Install dependencies:

   conda create -n visual-slam python=3.7
   conda activate visual-slam
   pip install -r requirements.txt
   

3. Run motion2D2D.py

And you're good to go!

Contact

• Hussein Lezzaik : hussein dot lezzaik at gmail dot com