
Document généré le 19/09/2025 depuis l'adresse: https://www.documentation.eauetbiodiversite.fr/fr/notice/stereovision-locale-et-reconstruction-3d-4d
Permalien: https://www.documentation.eauetbiodiversite.fr/fr/notice/stereovision-locale-et-reconstruction-3d-4d
Titre alternatif
Producteur
Contributeur(s)
Nice Sophia-Antipolis
Identifiant documentaire
9-6478
Identifiant OAI
oai:archimer.ifremer.fr:6478
Auteur(s):
Brandou, Vincent
Mots clés
3D reconstruction
camera trajectory
visual servoing
stereovision system
computer vision
3D metrology
reconstruction 3D
trajectoire d'acquisition
asservissement visuel
stéréovision
vision par ordinateur
métrologie 3D
Date de publication
09/12/2008
Date de création
Date de modification
Date d'acceptation du document
Date de dépôt légal
Langue
fre
Thème
Type de ressource
Source
Droits de réutilisation
info:eu-repo/semantics/openAccess
Région
Département
Commune
Description
The aim of this study is to propose a complete 3-dimension reconstruction method of natural submarine objects improved by a new acquisition method for quantitative measures, which can be used in operational conditions. First, it was necessary to take into account the various problems connected with the deep sea environment; the main constraint is that the system used to collect images must be manipulated at very important depths, up to 6000 meters by an underwater vehicle positioned on the sea floor. Thus, a method allowing the automatic acquisition of images was developed, adapted to any type of small-scale submarine object (approximately 1m3). The image acquisition is performed with a stereovision system operated by a manipulator arm. The method that we propose enables us to know extrinsic camera parameters by following a specific trajectory defined by the geometry of a stereo rig. Indeed, the trajectory is generated by the displacement of one camera onto the position of the other one by visual servoing. With this method, we can register images at regular intervals directly linked to the geometry of the stereo rig. Then, the 3D model of the underwater object is calculated from the collected images and camera parameters. The final result is a dense 3D reconstruction with texture mapping that enables metric measures.
Accès aux documents
0
Consultations
0
Téléchargements