Simple finite element-based computation of distance functions in unstructured grids
A distance field is a representation of the closest distance from a point to a given surface. Distance fields are widely used in applications ranging from computer vision, physics and computer graphics and have been the subject of research of many authors in the last decade. Most of the methods for...
Saved in:
Published in | International journal for numerical methods in engineering Vol. 72; no. 9; pp. 1095 - 1110 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
Chichester, UK
John Wiley & Sons, Ltd
26.11.2007
Wiley |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | A distance field is a representation of the closest distance from a point to a given surface. Distance fields are widely used in applications ranging from computer vision, physics and computer graphics and have been the subject of research of many authors in the last decade. Most of the methods for computing distance fields are devoted to Cartesian grids while little attention has been paid to unstructured grids. Finite element methods are well known for their ability to deal with partial differential equations in unstructured grids. Therefore, we propose an extension of the fast marching method for computing a distance field in a finite element context employing the element interpolation to hold the Eikonal property (∥∇φ∥ = 1). A simple algorithm to develop the computations is also presented and its efficiency demonstrated through various unstructured grid examples. We observed that the presented algorithm has processing times proportional to the number of mesh nodes. Copyright © 2007 John Wiley & Sons, Ltd. |
---|---|
Bibliography: | Center for Parallel Computations (NACAD) Petroleum National Agency (ANP, Brazil) Brazilian Council of Technological and Scientific Development-CNPq - No. CT-PETRO/PROSET 500.196/02-8 istex:FC8ACE09FE29ACA1767E814C4A9465A1612E9172 ArticleID:NME2079 ark:/67375/WNG-G0ZL4985-5 ObjectType-Article-2 SourceType-Scholarly Journals-1 ObjectType-Feature-1 content type line 23 |
ISSN: | 0029-5981 1097-0207 |
DOI: | 10.1002/nme.2079 |