Skip to Main content Skip to Navigation

Numérisation 3D intelligente d'objets de formes inconnues basée sur des critères de qualité

Abstract : Research works exposed in this thesis are devoted to 3D digitalization of unknown objects by laser plane sensor. We propose a new approach of intelligent automated scanning (IS) trajectory based on a quality criterion. Parameters allowing the relative positioning sensor/part as well as the consequence of their variation on the data point qu ality are studied. An experimental protocol, dedicated to performance assessment of the digitizing system is proposed. As a result, the system's capacities regarding to digitizing noise & the admissible space for the generation of IS trajectory are identified. The protocol is supplemented by assessment's procedures of the accuracy, the relative accuracy and fidelity. The approach proposed for IS of unknown object relies on a first scanning of the surfaces. The quality of the resulting points cloud is evaluated taking advantages of the previous step's result. The unsatisfied quality's zones (in ternis of digitizing noise and completeness) are identified. For those zones, the IS trajectory is generated to increase the quality. The new parameters (angle and distance of digitizing) provide a sensor position series, belonging to the admissible space previously defined. The whole approach is illustrated through several examples.
Complete list of metadata

Cited literature [83 references]  Display  Hide  Download
Contributor : Ens Cachan Bibliothèque <>
Submitted on : Wednesday, December 19, 2007 - 3:50:50 PM
Last modification on : Monday, February 15, 2021 - 10:39:37 AM
Long-term archiving on: : Friday, November 25, 2016 - 6:51:51 PM


  • HAL Id : tel-00199739, version 1


Charyar Mehdi-Souzani. Numérisation 3D intelligente d'objets de formes inconnues basée sur des critères de qualité. Mécanique []. École normale supérieure de Cachan - ENS Cachan, 2006. Français. ⟨tel-00199739⟩



Record views


Files downloads