摘要
提出一种三维散乱点云边界特征提取算法,该算法采用R^(*)-tree建立散乱点云的空间索引结构,基于该结构快速获取任意点的k近邻点集X,计算该点集的型心,依据型心及任意点构建向量v,建立经过点P且垂直于v的基准平面L,若点集X中各点均位于平面L的同侧,则点P为边界点,否则为非边界点。实例表明该算法运行速度快,且在快速准确提取点云边框特征点外,可同时提取孔洞处的边界特征点。
A new boundary extraction algorithm of unorganized point clouds is proposed in this paper.The topology of unorganized point clouds is constructed with R^(*)-tree,then the k-nearest neighborhood of the sample point is obtained,the core point of the k-nearest neighborhood is calculated,and the vector v is calculated which go through sample point and core point,then the datum plane is established which pass through sample point and perpendicular to vector.If all the k-nearest neighborhood points distributed on the same side of the datum plane,the sample point can be considered as boundary point,otherwise,it can be considered as non-boundary point.The implementation show that beside the outside boundary points,the hole's boundary points can also be extracted by the algorithm at the same time.
出处
《工业控制计算机》
2021年第4期116-117,共2页
Industrial Control Computer
基金
国家自然科学基金资助项目(51605037)
滨州学院航空专项项目(BZXYLG2008)。
关键词
散乱点云
k近邻点集
基准平面
边界特征提取
unorganized point clouds
K-Nearest Neighborhood
datum plane
boundary extraction
