摘要：

提出一种由截面上散点生成的最短路径拟合轮廓曲面的方法，生成二维轮廓点序列，根据各层轮廓拓扑上的相似性进行曲面重构，并进一步根据轮廓面重构进行曲面光滑。整个算法模拟了人在理解三维数据的过程，算法结果在截面上反映了散点数据总体走向，三维重构的曲面的形状细节保持较好。算法可以自适应处理截面上的联通数据，对噪声较大的数据鲁棒性较好。在三维地震体数据可视化的具体应用实验中获得较好的效果。

关键词: 轮廓曲面拟合, 最短连通路径法, 三维Delaunay剖分, 曲面平滑

Abstract:

This paper proposes a new way to reconstruct countour surface from unorganized points, which uses the shortest path method on the cross-sections. This technique generates the vertexes contour lists on each planes, then reconstructs the fitting surface according to the similarity of countours’ topology, and smooths the fitting surface. The algorithm simulates the process of human beings to understand the 3D data, and the result of the method shows that the details of 3D shape are maintained well. And the algorithm can handle the connected data points on the cross-section, and keep robustness on noised data. The method works well in 3D seismic data visualization.

Key words: contour surface fitting, the shortest path method, 3D Delaunay part, surface smoothing

中图分类号: 

• N945