The invention provides a method to estimate the scattered point cloud based on the neighborhood isomorphic surface of the sample point, which belongs to the field of product digital design and manufacture. By rebuilding the neighborhood point set, we get the local mesh surface with topological isomorphism on the original surface, extract the neighborhood area of the sample points in the grid surface, and estimate the normal direction of the sample point based on the neighborhood surface normal direction. In order to achieve smooth transition from point cloud to normal direction, Gauss kernel function is used to determine the weight coefficients of different neighborhood points of sample points, and the normal direction of samples is corrected by weight calculation. This method has high computational efficiency, and has a reliable estimation effect in the sharp feature area, and has a certain inhibitory effect on the noise.
【技术实现步骤摘要】
基于样点邻域同构曲面的散乱点云法向估计方法
本专利技术提供一种基于样点邻域同构曲面的散乱点云法向估计方法,属于产品数字化设计与制造领域。
技术介绍
点云数据具有简洁、灵活,并且无需保持拓扑一致等特性,在工业制造、文物保护及医学等领域得到广泛的应用。法向作为点云的重要属性之一,其计算结果的准确性将直接影响点云数据后期处理效果。目前法向估计主要可分为基于局部样本的方法和依据点云全局信息的方法两类。Hoppe等在《Surfacereconstructionfromunorganizedpoints》(ACM,1992)中采用局部表面拟合的方法,在假定采样表面平滑的前提下,利用样点局部k邻域点集进行平面拟合,以所得平面的法向作为样点法向。拟合的方法原理简单,但计算精确度较低,当采样不均匀或模型结构复杂时,其法向估计结果往往误差较大。Gouraud等《Continuousshadingofcurvedsurfaces》(IEEEtransactionsoncomputers,1971,100(6):623-629.)中为样点构建邻域三角面片,将样点的法向定义为所构建三角面片 ...
【技术保护点】
一种基于样点邻域同构曲面的散乱点云法向估计方法,其特征在于对S中任一样点pi,其法向估计步骤依次为:(1)求取pi的邻域点集并进行局部重建,获得局部重建网格
【技术特征摘要】
1.一种基于样点邻域同构曲面的散乱点云法向估计方法,其特征在于对S中任一样点pi,其法向估计步骤依次为:(1)求取pi的邻域点集并进行局部重建,获得局部重建网格(2)依据中面片法向初步估计邻域样点法向;(3)采用Dijkstra算法计算目标样点pi至邻域点集中任一样点的近似测地距离,并基于高斯核函数求解pi法向;(4)当对S中所有样点遍历结...
【专利技术属性】
技术研发人员:孙殿柱,薄志成,李延瑞,梁增凯,
申请(专利权)人:山东理工大学,
类型:发明
国别省市:山东,37
还没有人留言评论。发表了对其他浏览者有用的留言会获得科技券。