Original-Design-Intention-Driven Parametric Surface Reconstruction Method

The conventional surface reconstruction methods and software can create a similar surface model with physical model by directly fitting faces to corresponding point cloud data and combining them. But for some special application which requires a consistency 3D parametric model with original design logic and engineering rules, previous methods are very inefficient or even impossible. This paper presents a new design-intention-driven surface reconstruction method to solve this kind of problems. A multi-stage optimization mechanism basing on sensitivity analysis will optimize the many design variables in this imprecise model driven by original design intentions and minimize the deviation between model and point cloud to a level of manufacturing tolerance. And then the engineers obtain the consistency model as well as the key design parameters for improving design, deriving new products and later processing. This paper uses the raceway surface reconstruction of Birfield ball-joint as an example to illustrate this method.