Direct digital design and manufacturing from massive point-cloud data

Scanned point-cloud data plays an increasingly important role in various applications such as shape modeling, graphical rendering and engineering design and manufacturing. On the one hand, various 3D sensors can now readily output dense and accurate point cloud of a physical object. On the other hand, the use of geometric models of existing physical objects is becoming increasingly common in the product development.;We present an approach that can enable product geometric design and its subsequent manufacturing applications directly from massive point-cloud data, without the usual laborious CAD model reconstruction. We term this approach direct digital design and manufacturing (D3M).;In this approach, we adopt the projection based moving least-squares (MLS) surface as the single underlying geometry presentation, which defines a continuous surface directly from a set of points. Based on this, we perform differential geometric and topological analysis for point-set surfaces: in differential geometric analysis, we derive analytical formulae for curvature computing for MLS surfaces/curves; in differential topological analysis, we develop a new method to generate Morse-Smale complex for a continuous function restricted to the implicitly defined MILS surface.;The above geometric and topological analysis enables the development of efficient (curvature-adaptive), accurate (error-bounded) and robust (topologically guaranteed) surface intersection algorithms that intersect the MILS surface with lines, planes, polygonal meshes, and NURBS surfaces.;These algorithms have been successfully applied in applications ranging from direct shape design to direct manufacturing: in direct shape design, we have developed a novel direct shape modeling method based on both designed geometry and acquired geometry; In direct manufacturing, we have presented MLS based approaches for direct slicing in rapid prototyping and path generation in NC machining. In comparison with traditional procedures, our D3M approach avoids the troublesome surface reconstruction and the potential model conversion induced accuracy loss.;In conclusion, a D3M approach has been developed for product development involving scanned point-cloud data. Based on MLS as a valuable surface representation for acquired point-sampled geometry, differential geometric and topological analysis have been given and proved to be useful for D3M algorithms and applications.