大规模孔洞点云的快速重建算法研究 *

针对实际中经常存在的含有孔洞的点云数据 ,在原多层重建算法的基础上提出了一种可以进行点云补洞的快速曲面重建算法。首先对散乱点云数据进行空间自适应八叉剖分 ,然后对点云数据进行由粗到精的多层插值 ,建立隐式曲面方程 ,最后提出了两种加快重建的方法。加速算法可以减少重建时间 ,非常有利于处理大规模点云。实验结果证明 ,本算法对点云孔洞修补效果良好 ,重建速度快 ,效率高。

Algorithmic research on fast reconstruction from large point clouds with holes

For point clouds with holes often found in practice, developed a fast surface reconstruction algorithm, which could repair incomplete point data. First, divided scattered point data adaptive into space with octree, Then interpolated point data by a coarse-to-fine hierarchy, to get implicit surface equation. Finally, introduced two optional methods which could make o-riginal algorithm faster. The accelerating algorithm is very useful for large point clouds, which can shorten reconstruction time. The experimental results demonstrate the algorithm is fast, efficient and effective.