针对非密度均匀的点云,提出了一种高效保持特征的曲面重建算法。首先利用八叉树进行点云空间分割,然后对每个点在小邻域内求出局部逼近曲面,建立隐式曲面方程。通过参数限定点的邻域范围,使整个算法既保证了重建效果,又不致于很大程度上增加重建时间,达到了速度和效果在一个范围内的平衡。实验结果证明,本算法重建效果良好,适用于各种散乱点云的重建。
For non-uniform point clouds,this paper presented a new algorithm for surface reconstruction. It carried out octree to divide the point clouds into space, computed local approximation in a small vicinity for each point, then got the implicit equation. With the parameter limiting the scope of the neighborhood of point, the algorithm was to ensure the effectiveness of the reconstruction and would not increase much time to a large extent, with a balance of the speed and effectiveness. Experimental resuhs show that the reconstruction algorithm is effective and robust.