pyslam

This project demonstrates a simple implementation of a monocular SLAM system using ORB feature detectors

Features

MP4 video loader
KITTI Odometry dataset loader
Point cloud visualizer

Requirements

Python 3.7+
numpy
OpenCV
g2o
pangolin

For g2o and pangolin python bindings, use the forks in my github.

Usage

For video files chmod +x system.py && ./system.py --type <VIDEO> --path <test.mp4>
For Kitti dataset chmod +x system.py && ./system.py --type <KITTI> --path </path/to/kitti/dataset/sequences/00>

License

All of this code is MIT licensed. Videos and libraries follow their respective licenses.