VAR-SLAM: Visual Adaptive and Robust SLAM for Dynamic Environments

VAR-SLAM introduces a novel feature-based visual SLAM system designed for dynamic environments. It combines a lightweight semantic keypoint filter to deal with known moving objects, with Barron’s adaptive robust loss to handle unknown ones. An adaptive parameter α is estimated online, allowing the system to handle both static and dynamic regions effectively without extensive threshold tuning.

[Paper] [arXiv] [Project page]

If you use VAR-SLAM in your work, please cite the following paper:

João Carlos Virgolino Soares, Gabriel Fischer Abati, Claudio Semini, "VAR-SLAM: Visual Adaptive and Robust SLAM for Dynamic Environments", arXiv preprint arXiv:2510.16205, 2025.

@article{Soares2025VARSLAM,
      title={VAR-SLAM: Visual Adaptive and Robust SLAM for Dynamic Environments},
      author={João Carlos Virgolino Soares and Gabriel Fischer Abati and Claudio Semini},
      year={2025},
      eprint={2510.16205},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2510.16205}
}

Code

Code coming soon.

Acknowledgements

Our code builds on ORB-SLAM3.

Maintainer

This repository is maintained by João Soares and Gabriel Fischer.