Issue |
ESAIM: Proc.
Volume 27, 2009
CANUM 2008
|
|
---|---|---|
Page(s) | 254 - 271 | |
DOI | https://doi.org/10.1051/proc/2009031 | |
Published online | 25 June 2009 |
Level set driven smooth curve approximation from unorganized or noisy point set
1
UPMC Univ Paris 06, UMR 7598, Laboratoire J.-L. Lions, F-75005 Paris, France.
2
Universidad de Chile, UMI 2807, Centro de Modelamiento Matemático, Santiago, Chile.
Corresponding authors: claisse@ann.jussieu.fr frey@ann.jussieu.fr
In this paper, we propose a curve construction method for a non uniform point data set based on a minimal curve approximation model. Numerically, the level set method is used for curve reconstruction. We represent the shape of the curve through its distance function and formulate curve reconstruction as a constrained minimization problem. We solve the minimization problem on a highly anisotropic triangulation to improve the accuracy of the numerical scheme. This method can handle complex geometries and deal with arbitrary topologies as well as with noisy data sets. Several numerical examples are provided to show the efficiency of the proposed approach.
Résumé
Dans ce papier, on propose un modèle de courbe d'approximation minimale pour construire une courbe à partir d'un nuage de points. Numériquement, la reconstruction de la courbe s'appuie sur une formulation de type ligne de niveau. On représente la forme de la courbe par sa fonction distance aux points et on exprime ce problème comme un problème de minimisation. Ce dernier est résolu sur une triangulation anisotrope qui permet d'améliorer la précision du schéma numérique. Cette méthode permet de traiter des géométries complexes et des topologies quelconques ainsi que des données bruitées. Des exemples de reconstruction sont proposés pour montrer l'éfficacité de cette approche.
© EDP Sciences, ESAIM, 2009
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.