Skip to Main content Skip to Navigation
Theses

Fusion de données multicapteurs pour la construction incrémentale du modèle tridimensionnel texturé d'un environnement intérieur par un robot mobile

Abstract : This thesis examines the problem of 3D Modelling of indoor environment (supposed unknown) by a mobile robot. Our main contribution consists in constructing a dense geometrical model represented as a heterogeneous map containing textured planar landmarks, 3D lines and interest points. In order to achieve this mission, we must fuse geometrical and photometrical data. Hence, we began by improving the stereo vision algorithm. We proposed a new approach that transform the stereo matching into a problem of minimization of a global energy function. The minimum of this function is found by finding a minimum cut in a graph. The most significant contribution on stereovision is to propose, for the first time, the construction of a reduced graph that allows to accelerate the stereo correspondence using graph cuts and to provide better results than the local methods. Nevertheless, this method remains non applicable in mobile robotics. Also, to perceive the environment, the robot is equipped by a laser scanner rotating around a horizontal axis and by a camera. We proposed an algorithmic chain allowing to construct incrementally the heterogeneous map, using the algorithm of Simultaneous Localization and Mapping based on the extended Kalman filter (EKF-SLAM). Mapping the texture on the planar landmarks makes the model more realistic for a user, and makes more robust the phase of data association, which is essential to guarantee the map consistency. The different developed mechanisms were illustrated by experimental results.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00340911
Contributor : Arlette Evrard <>
Submitted on : Monday, November 24, 2008 - 8:35:11 AM
Last modification on : Friday, January 10, 2020 - 9:08:08 PM

Identifiers

  • HAL Id : tel-00340911, version 1

Citation

Ayman Zureiki. Fusion de données multicapteurs pour la construction incrémentale du modèle tridimensionnel texturé d'un environnement intérieur par un robot mobile. Automatique / Robotique. INSA de Toulouse, 2008. Français. ⟨tel-00340911⟩

Share

Metrics

Record views

389

Files downloads

364