摘要
针对现有法向量估值算法都只能适用于某一类特定形状模型的问题,提出三维点云模糊分类的法向量估值算法,利用模糊推理系统对模型的点云数据分类,根据点云在不同形状区域的分布情况和曲率变化给出模糊规则,将点云分成属于平滑形状区域、薄片形状区域和尖锐形状区域三类,每类点云对应给出特定的法向量估值算法.由于任意模型形状分布的差别,其点云数据经过模糊分类后调用相应的估值算法次数会有差别,因此采用牙齿模型点云数据验证了算法的可行性,经过与三种典型算法比较可以看出本算法估算准确、简单可行.
A normal estimation algorithm could be used to any shape model, because existing normal estimation algorithms were only suitable to some shape models. The point clouds of the model were classified by fuzzy inference systems. The fuzzy rulers were given according to the distribute characters and curvature change of point clouds. Three-dimensional point clouds were divided three categories, including smooth, thin and sharp shapes. The normal estimation algorithm of every kind of point was given. Because the shape distribution of the model was different, the called numbers of corresponding normal estimation algorithm were different after classifying too. The denture model was used to evident the algorithm. Compared with three typical algorithms, the algorithm can estimate the normal and construct the denture more accurate, efficient and simple.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第8期50-54,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(51205093
50675054)
关键词
三维点云
模糊推理
法向量估值
平滑形状
尖锐形状
薄片形状
three-dimensional point clouds
fuzzy inference
normal estimation
smooth shape
sharp shape
thin shape
