基于SOFM和NURBS的散乱数据点自由曲面重构

提出了利用NURBS实现曲面散乱点数据几何建模的方法。首先利用Unigraphics软件设计原始曲面，在所设计的原始曲面上划分若干条曲线，并在曲线上随机设置大量的散乱点数据；然后用自组织神经网络SOFM对原始曲面上的散乱数据点进行聚类，构建NURBS曲线并完成曲面重构。最后通过仿真实验分析了重构曲面相对原始曲面的误差，实验表明所提出的几何建模方法是有效的，并能较精确拟合原始曲面。

Free form surface reconstruction from scattered points based on SOFM and NURBS

A geometric modeling method, which based on NURBS from scattered points, was presented. By using Unigraphics the original surface was built, on which several curves were divided and a large number of scattered points on the curves were generated stochastically. By the clustering of scattered points with SOFM （Self-Organizing Feature Map）, the NURBS were built and the free-form surface reconstruction was accomplished. The error between the reconstructed surface and the original surface was analyzed through simulation. Results showed that the proposed method could reconstruct curved surface efficiently which could fit the original surface accurately.