Calage robuste et accéléré de nuages de points en environnements naturels via l'apprentissage automatique
|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. For this purpose, landmarks called 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. The presence of these unreliable descriptors adversely affects the performances of the alignment process. Therefore, we propose to filter 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.|
|Document Type:||Mémoire de maîtrise|
|Open Access Date:||19 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.