A surface is often approximated by a network of triangular facets. In the absence of a precise mathematical description of the underlying surface all information about surface properties such as smoothness and curvature must be inferred from the triangulation itself. Enrichment and coarsening of the surface geometry, for example, can only be carried out if singular features, where C1 continity is lost, are properly accounted for. In the absence of a well defined surface geometry it is necessary to extract these features and establish suitable data structures so that the features persist after enrichment and/or coarsening of the triangulation. This paper describes a feature extraction scheme that is based on estimates of the local normals and principal curvatures at each mesh point. The feature extraction scheme has been combined with an algorithm that adapts a tetrahedral mesh by the selective enrichment and coarsening of both the volume and surface triangulation.
Timothy J. Baker