一种基于立方格的散乱点云曲面网格化方法

A Cube Based Surface Meshing Method from Unstructured Points Set

提出一种新的基于小立方格的网格化方法。首先将空间划分为小立方格,计算小立方格位于物体表面的可信度,然后对可信度高的小立方格进行重采样,最后依据小立方格的空间邻接关系连接重采样点,形成三角网格。与Marching Cube方法相比较,该方法不需要考虑等值曲面(iso-surface),不需要计算小立方格8个顶点的符号,通过法向一致性区分物体表面的内外。实验表明该方法适用于拓扑变化复杂的散乱点云的景物建模,效果良好。

The authors give a new meshing method which is based on cubes. Firstly it divides the space into cubes, and calculates the confidence of the cubes that are on the object surface. Secondly it computes the resample vertices at the cubes with high confidence. Thirdly it links the resample vertices according to the cubes index and forms triangles mesh. The new meshing method does not need to consider iso-surface and inside/outside of the object, and does not need to compute the sign values at the eight corners of the cubes and distinguish the surface sides according to the normal consistency. The experiment shows that new method is suitable for unstructured point set with high topology change, and it is effective.