SIMULATION BASED MODELLING
Hans Hagen, Siegfried Heinz, Michael Thesing, THOMAS SCHREIBER · International Journal of Shape Modeling · 1998
The generation of a 3D CAD model from a set of unstructured points of an unknown 3D object is an important problem of a reverse engineering process that can occur several times within a complete design process chain. Efficient 3D scanning technologies, such as laser range scanning systems become more and more sophisticated. Therefore the representation of 3D objects as a large set of digitized, unstructured point data became a common technique that builds the first step of a reverse engineering process. The main interest of this paper is to explain the further steps of the reverse engineering process, the filtering, clustering and segmentation of the point data and the final surface reconstruction out of point segmentations.This paper is concerned with structuring of point data as a pre-processing before the actual surface reconstruction and with the final surface reconstruction based on a variational design approach considering boundary conditions. First, the focus is set to an algorithm to reduce the data of large point sets. Subsequent a new method for curvature approximation, based on a Delaunay triangulation of the points, is described. By means of curvature discontinuities the points are grouped to represent separate surfaces of the model. After that variational design methods considering boundary conditions to generate the final surfaces from the group of points tangent to earlier created surfaces will be explained.