Simple implementation of monocular visual odometry with loop closure detection.

Rust version here: https://github.com/markoelez/minislam2.

• ORB feature detection and matching
• Essential matrix decomposition for pose estimation
• Loop closure detection - detects when revisiting previously seen locations
• Real-time 3D visualization with loop closure visualization (yellow lines)

run the following for a quick example:

uv run main

using your own data:

1. start by doing one of the following:
   • download a sequence of images: https://www.cvlibs.net/datasets/kitti/raw_data.php
   • download an MP4 video
2. move the dataset into the eval/ directory
3. add dataset metadata to the config.yaml file
4. run with uv run main --dataset=[your_dataset_name]