Surface Parameterization And Reconstruction Of Automotive Seat Point Cloud Data Based On B-Spline

With the development of times, as a symbol of the economic globalization, automobile industry develops rapidly in worldwide. Because of the importance to the cars; the design and production of car seats obtain such a good opportunity to high-speed development correspondingly. However the traditional design pattern that accords to the design process by using the CAD software or other analogous software can not adapt to the current market which aims to reduce the time of design. As a kind of new design pattern which has higher efficiency and lower cost, reverse engineering develops rapidly under the demand of the design market. Therefore it has great theoretical value and practical value by using reverse engineering technology to design the automotive seats efficiently. That is the background of this paper.Reverse engineering mainly contains several parts, such as the point cloud data measurement, the point cloud data pretreatment, surface reconstruction and surface parameterization. This paper introduces and researches these key technologies in main.The original point cloud data is a kind of scan-line point cloud data which obtained by using laser scanning measuring equipments. Because of the high precision of scanning, it produced a large amount of data, which also naturally included a significant proportion of redundant data. So reducing the original point cloud data is the first step of the point cloud data pretreatment. We can make the point cloud into stratification in the scan-line direction, and then using the method which we call Sting tall-Angle deviation combining algorithm to reduce the data in 2D plane. Stratification can not only simplify 3D data of topological relationships but also reduce the redundant data between the parallel planes. And Sting tall-Angle deviation combining algorithm is not only in conformity with data distribution characteristics, but also good to keep curvature characteristics of original point cloud data. Usually in the reverse engineering design, the original surfaces of model are complex. We need combine a number of small curved surfaces to recreate the surface of original model. In this paper we roughly segment the point cloud data into several blocks manually. And then we segment the point cloud data automatically by extracting feature points which bases on the calculation of the surface curvature.There are two ways to reconstruct the four-boundary surfaces usually. In this paper we reversely calculate the control points with the value points first, then reconstruct the curved surface. We could fit out the value point grid with the method which designed by my own, which we call local bounding box of K neighborhood weighted average method. Next, we set the value points on the surface boundary as the control points. We set the simultaneous equations with the relationship between the control points and the value points of dual three times uniform B-Spline curve, and then reversely calculate the control point grid. After that we render the surface according to the expression of dual three times uniform B-Spline curve. And then we make the adjacent surface edges transited smoothly by using the interpolation method. Finally we form integrated model.Specifically we can draw the control point grid in the surface reconstruction process. We can interactively modify the control points by picking and moving with the mouse and keyboard. We reconstruct and render the surface according to the modified parameters to realize the parameterization of curved surface design.In the last part, we design a method of surface reconstruction which lofts the data points between the data layers along the central axis.