toy_vo

A very simple stereo visual odometry library.

Run Example

You can download datasets from

• EuRoC MAV Dataset
• TUM Visual-Inertial Dataset
• TUM 4Seasons Dataset

Python

pip install toy-vo rerun-sdk==0.32 opencv-python scipy
python3 examples/run_stereo.py -d {your_path/V1_01_easy} -c configs/euroc --rerun

Rust

git clone https://github.com/powei-lin/toy_vo.git && cd toy_vo
cargo run -r --example run_stereo -- -d {your_path/dataset-corridor4_512_16} -c configs/tum_vi --rerun

Calibrate your own camera and run toy-vo

Use camera-intrinsic-calibration and format your images like euroc dataset.

Acknowledgements

Thanks to all the authors of these libraries.

• basalt
• faer
• nalgebra
• rerun

And thanks to Hossam R. for implementing the LM method in tiny-solver.