Détail de l'auteur
Auteur J. Guhring
Documents disponibles écrits par cet auteur
Ajouter le résultat dans votre panier Affiner la recherche Interroger des sources externes
3D-Erfassung und Objektrekonstruktion mittels Streifenprojektion / J. Guhring (2002)
Titre : 3D-Erfassung und Objektrekonstruktion mittels Streifenprojektion Titre original : [Saisie 3D et reconstruction d'objet au moyen de la projection par bandes] Type de document : Thèse/HDR Auteurs : J. Guhring, Auteur Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2002 Collection : DGK - C Sous-collection : Dissertationen num. 560 Importance : 166 p. Format : 21 x 30 cm ISBN/ISSN/EAN : 978-3-7696-9599-1 Note générale : Bibliographie Langues : Allemand (ger) Descripteur : [Vedettes matières IGN] Photogrammétrie numérique
[Termes descripteurs IGN] méthode robuste
[Termes descripteurs IGN] modèle mathématique
[Termes descripteurs IGN] reconstruction 3D
Résumé : (Auteur) The digitization of physical objects using optical shape acquisition methods is gaining increasing importance. Nowadays, there are a lot of applications in many different fields like industrial inspection, reverse engineering, virtual reality and medicine. Due to the improvements in sensor architecture and data processing, the acquisition of objects of almost any complexity became possible.
Active research during the last decades led to a variety of working shape acquisition systems worldwide. They differ in the measurement principle, efficiency and the degree of automatization. What almost all systems have in common is the missing or inadequate consideration of accuracy and reliability aspects.
This thesis addresses the problem of surface acquisition and reconstruction using active optical triangulation. Compared to other systems, data quality is treated rigorously in all processing steps. In order to achieve this goal new methods or important extensions of existing methods have been developed. The necessary steps are :
Data acquisition using structured light. Calibration and modeling of cameras and projectors is achieved by adapting well-established methods of digital close range photogrammetry to fringe projection systems. New pattern designs in combination with advanced processing methods allow for higher flexibility without introducing systematic measurement errors, as they are inherent in traditional methods. Results of the process are the geometric positions of the measured points, as well as, their covariance matrix and additional attributes like color or surface reflectivity.
Transformation of datasets into a common coordinate frame. This so-called registration step is necessary since almost any measurement task requires multiple measurements from different viewpoints due to unavoidable occlusions. Algorithms for pairwise registration as well as algorithms for the simultaneous registration of multiple datasets and registration using surface reflectivity information have been investigated. All methods are modified versions of the well-known iterative closest point (ICP) algorithm and make use of covariance information provided by the data acquisition step. This information is used during the pose estimation process. A new stochastic model, which takes into consideration how point correspondences are computed, tremendously improves convergence compared to the classical ICP algorithm.
The integration of registered datasets. This step modifies the topology of the measured points to yield a seamlessly integrated surface description.
For measurement tasks in industrial inspection, the deviation between a manufactured part and its CAD description is necessary (actual/nominal comparison). The uncertainty of the reconstructed surface is used to decide whether a measured deviation is significant and if the measurements are precise enough to detect small defects.
The realized system allows for the digitization of complex objects using millions of points. By the integration of accuracy and reliability aspects in all steps of the processing chain the robustness of the system has been considerably improved.
Numéro de notice : 13162 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Permalink :
RéservationRéserver ce document
Code-barres Cote Support Localisation Section Disponibilité 13162-01 33.30 Livre Centre de documentation Photogrammétrie - Lasergrammétrie Disponible