Surface Digitization And Nurbs-based Surface Reconstruction

Reverse engineering is an important research field in mechanical engineering. And as a vital method of rapid manufacture, reverse engineering is taken into account widely in the field of CAD/CAM. Especially, as an important process in reverse engineering, surface reconstruction is always attached importance to by CAD/CAM Software Company and will catch more attention in future work.With the analysis on the currently used method and research fruit, this paper present a new method for surface reconstruction according to the process in surface reconstruction.The theoretical part deals with the theory about surface digitizing, and researched the widely used 3D laser scan principle. Research the theory and method on surface reconstruction from a topologically arbitrary, data points cloud a representation of this surface to a set of smoothly joining NURBS patches. The first start the computation of a triangular surface from this point set based on the surface reconstruction method of Hoppe H., then process the conversion of the triangular surface to quadrangular surface according to Catmull-Clark subdivide algorithm and the last build the conversion of this quadrangular surface to a NURBS surface achieved by using the PCCM algorithm of J. Peter.The practical part describes the reconstruction pipeline that has been designed and developed according to software engineering and design patterns, which could make the software be used to research the new algorithm and technology. Researched the half-edge data structure, implement the structure to organize 3D mesh. Researched track ball algorithm, implement the method to operate 3D objects in 2D interface. The different models in the developed system could be assembled through data stream and file stream. The surface reconstruction software developed under Visual Studio .Net 2003.Digital simulation and measuration of real object were adopted to check the surface reconstruction method. The results show that the method is simple and easy, the reconstructed surface reached the precision at some extent. The method is fit for processing surface reconstruction from the dense data points.