Skip to Main content Skip to Navigation
New interface
Journal articles

Reconstruction de surfaces B-splines régularisées et tridimensionnelles à partir de contours d'occultation

Changsheng Zhao 1 Roger Mohr 1, 2 Boubakeur Boufama 1, 2 
2 MOVI - Modeling, localization, recognition and interpretation in computer vision
GRAVIR - IMAG - Laboratoire d'informatique GRAphique, VIsion et Robotique de Grenoble, Inria Grenoble - Rhône-Alpes, CNRS - Centre National de la Recherche Scientifique : FR71
Abstract : The three dimensional surface reconstruction of a non polyhedral object is a difficult problem in computer vision . In this paper, a new method for reconstructing three dimensional surface from the recovered motion of occluding contours is presented through calibrated image sequences . We use the uniform bicubic Bspline surface patches to give a parametric representation of an object surface . Finally, the problem of three dimensional B-spline surface patches reconstruction is equivalent to find their control points by solving a nonlinear system . Two numerical methods are outlined : Levenberg-Marquardt, Quasi-Newton . To avoid the classic camera calibration that needs a calibration pattern, we propose a direct nonlinear method of the autocalibration of a camera using the stable points in the scene. Our approach can be applied in the case where the camera is calibrated, the object is smooth, specifically, that its surface is at least C2. Results for reconstruction based on synthetic and real data are presented .
Document type :
Journal articles
Complete list of metadata
Contributor : THOTH Team Connect in order to contact the contributor
Submitted on : Monday, December 20, 2010 - 8:44:21 AM
Last modification on : Friday, February 4, 2022 - 3:23:47 AM


  • HAL Id : inria-00548395, version 1



Changsheng Zhao, Roger Mohr, Boubakeur Boufama. Reconstruction de surfaces B-splines régularisées et tridimensionnelles à partir de contours d'occultation. Traitement du Signal, 1995, 12 (2), pp.129--143. ⟨inria-00548395⟩



Record views