基于轮廓约束点的B样条曲面拟合算法

提出了一种面向截面测量数据的B样条曲面拟合算法。首先对原始数据点列进行降噪处理，然后遴选出曲率优势点，并将其作为初始的轮廓约束点，得到插值于约束点的初始曲线。再在需改善拟合精度的区域增加约束点，直至获得满足精度要求的B样条曲线。最后以约束点数目最多的曲线为准，在其余的曲线上增加差额数目的约束点，并进行平均弦长参数化，构造出B样条曲线簇，最终获得B样条拟合曲面。仿真实验结果表明，该方法可显著压缩曲面模型的控制顶点数目，具有较高的曲面重构效率。

B-spline Surface Fitting Algorithm Based on Contour's Constraint Points

Aiming at cross-sectional measured data, a B-spline surface fitting method was proposed. The original data points were denoised firstly, then vantage points of curvature were picked out and taken as the initial contour constraint points, and an initial curve interpolated with the constraint points was gained. After that, constraint points were added to the region whose fitting accuracy needed to be improved, until the B-spline curve could reach preset accuracy. Finally, subjecting to curve which contained the most constraint points, constraint points of the balance number were added to the rest of the curve, and average chord-length parameterization was conducted to construct a family of B-spline curves, and a B-spline fitting surface was obtained at last. Simulation results show that, this method can compress the number of control vertices of the surface model significantly with a high efficiency of surface reconstruction.