基于八叉树的海量测量数据的非均匀简化

提出了一种高效的非均匀简化方法：先求出数据集的最小立方体包围盒，应用八叉树原理将其分割成八个小立方体；然后计算每个非空立方体中局部离散曲面的曲面变分和包含的点数，并根据用户给定的曲面变分阈值和点数阈值决定是否继续分割；最后为每个叶子立方体保留一个点。应用实例表明，该方法速度快，能很好地保留原始数据的几何特征，适用于复杂的测量数据。

Nonuniform simplification of cloud data based on octree

An efficient nonuniform simplification algorithm was proposed. At first, the minimum cube box of the data sets was gained, and it was divided into eight cubes with employment of oetree. Then surface variation of local discrete surface and the number of points in nonempty cubes were calculated, and these cubes were divided continually or not according to the two user-specified thresholds. In the end, one point was reserved for each leaf cube. Practical examples show that the efficient method can exactly retain geometric characteristics of original dates and be applicable for complex cloud data.