Automatic building of complex triangulated surfaces: latest improvements

@inproceedings{LedezRM2001,
 abstract = { This paper describes the continuation of the research work presented last year on the development of new algorithms for building automatically (or at least as much as possible automatically) complex triangular surfaces from sets of data points. The methodology introduced here assumes the surface to be an iso-value of a function f (x,y,z) de£ned in the 3D space. The proposed algorithm consists of four main steps: (i) the initial set of points is partitioned in such a way that each part is as ¤at as possible. (ii) elementary patches are built from each subset to estimate the normal at each points of the cluster. (iii) each patch is then duplicated and translated according to its normal direction (upwards and downwards) in order to de£ned a function f (x,y,z) to the values -1 and +1 and on these two copies. (iiii) £nally the function f (x,y,z) is estimated everywhere in the 3D space; the surface is then extracted as the iso-value f (x,y,z) = 0. Improvements have been included in the new version of the algorithm, making the iso-value technique an ef£cient tool for building automatically a surface from an unstructured set of points. Furthermore, new data, namely cross-sections, have been taken into account by de£ning another implicit function. Some case studies show the ef£ciency of this method. },
 author = { Ledez, David AND Cognot, Richard AND Mallet, Jean-Laurent },
 booktitle = { 21st gOcad Meeting },
 month = { "june" },
 publisher = { ASGA },
 title = { Automatic building of complex triangulated surfaces: latest improvements },
 year = { 2001 }
}