Efficient CAD-based triangulation of unordered 3D data

Abstract. The paper proposes a new method for efficient triangulation of large, unordered sets of 3D points using a CAD model comprising NURBS entities. It is primarily aimed at engineering applications involving analysis and visualisation of measured data, such as inspection, where a model of the object in question is available. Registration of the data to the model is the necessary first step, enabling the triangulation to be efficiently performed in 2D, on the projections of the measured points onto the model entities. The derived connectivity is then applied to the original 3D data. Improvement of the generated 3D mesh is often necessary, involving mesh smoothing, constraint-based elimination of redundant triangles and merging of mesh patches. Examples involving random measurements on aerospace and automotive free-form components are presented.