Accelerated Robust Point Cloud Registration in Natural Environments through Positive and Unlabeled Learning

Latulippe, Maxime; Drouin, Alexandre; Giguère, Philippe; Laviolette, François

Accelerated Robust Point Cloud Registration in Natural Environments through Positive and Unlabeled Learning

Maxime Latulippe, Alexandre Drouin, Philippe Giguère, François Laviolette

Download BibTeX Cite

Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence. Pages 2480-2486.

Abstract

Localization of a mobile robot is crucial for autonomous navigation. Using laser scanners, this can be facilitated by the pairwise alignment of consecutive scans. In this paper, we are interested in improving this scan alignment in challenging natural environments. For this purpose, local descriptors are generally effective as they facilitate point matching. However, we show that in some natural environments, many of them are likely to be unreliable, which affects the accuracy and robustness of the results. Therefore, we propose to filter the unreliable descriptors as a prior step to alignment. Our approach uses a fast machine learning algorithm, trained on-the-fly under the positive and unlabeled learning paradigm without the need for human intervention. Our results show that the number of descriptors can be significantly reduced, while increasing the proportion of reliable ones, thus speeding up and improving the robustness of the scan alignment process.

Maxime Latulippe, Alexandre Drouin, Philippe Giguère, and François Laviolette. 2013. Accelerated Robust Point Cloud Registration in Natural Environments through Positive and Unlabeled Learning. In Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence, IJCAI-13. IJCAI, 2480–2486. https://ijcai.org/proceedings/2013/365

@inproceedings{ijcai2013p365,
  title     = {Accelerated Robust Point Cloud Registration in Natural Environments through Positive and Unlabeled Learning},
  author    = {Latulippe, Maxime and Drouin, Alexandre and Giguère, Philippe and Laviolette, François},
  booktitle = {Proceedings of the Twenty-Third International Joint Conference on
               Artificial Intelligence, {IJCAI-13}},
  publisher = {AAAI Press / International Joint Conferences on Artificial Intelligence},
  editor    = {Francesca Rossi},
  pages     = {2480--2486},
  year      = {2013},
  month     = {8},
  url       = {https://ijcai.org/proceedings/2013/365},
}

TY  - CPAPER
TI  - Accelerated Robust Point Cloud Registration in Natural Environments through Positive and Unlabeled Learning
AU  - Latulippe, Maxime
AU  - Drouin, Alexandre
AU  - Giguère, Philippe
AU  - Laviolette, François
T2  - Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence, IJCAI-13
PB  - AAAI Press / International Joint Conferences on Artificial Intelligence
PY  - 2013
SP  - 2480
EP  - 2486
UR  - https://ijcai.org/proceedings/2013/365
ER  -