Evolutionary surface reconstruction using CSG-NURBS-Hybrids

Abstract

The aim of evolutionary surface reconstruction is to find mathematical descriptions of surfaces of physical objects. These objects are represented only by discrete sampling points. The reconstructions have to render the original shape of the objects as realistic as possible. Therefore, complex discrete approximation problems must be solved. In order to process surface reconstructions in CAD systems, the descriptions have to render the construction logic of the scanned object. Here, Nonuniform Rational B-Splines (NURBS) and Constructive Solid Geometry (CSG) are combined by hybrid reconstruction system that fits each surface type automtically using one homogenous data structure.