基于重新检测过程的三维细化算法的改进

摘要

The existing thinning algorithms based on simple point fail to preserve the connectivity of extracted skeleton.This paper first proposed a set of isotropic deleting templates which keeps the algorithm have 90°rotation invariance,and then proposed a new re-checking procedure by detecting weather the connectivity of target point's 26-neighborhood have changed or not after deleting points to determine whether the target should be reduced,thus the connectnity of 3D objects can be detected by the point.This method can suit most of the thinning algorithms based on simple point and fix the cavities to preserve topology structure.The new proposed algorithm can get the best results when compared with other algorithms based on templates in rotation invariance test.%现有的基于简单点判断的三维细化算法不能保证提取骨架的连续性,容易产生断裂.针对该问题,提出了一组各向同性模板,该模板能够使得算法具有90°旋转不变性;在此基础上,进一步提出了一种新的重新检测的方法,通过判断被删除的目标点的26邻域的连通性,来决定该目标点是否应该被还原,从而逐点检测3D物体的连通性,达到保持整体连通性的目的.该方法可以应用于大多数基于模板的三维细化算法,能够修复断裂,保证其拓扑结构,避免产生空洞;同时,与同类算法相比,本算法由于利用了各向同性模板,在物体旋转的情况下亦能得到最佳的细化结果.