面向船体外板曲面重建的散乱点云精简算法  

Research on Streamline Algorithm for Ship Hull Plate Surface Reconstruction of Scattered Point Cloud

作　　者：孙志敬[1] 程良伦[2] 李晓娟[2] 

机构地区：[1]广东工业大学自动化学院,广东广州510006 [2]广东工业大学计算机学院,广东广州510006

出　　处：《机械设计与制造》2014年第9期228-231,共4页Machinery Design & Manufacture

基　　金：广东省教育部产学研结合项目(2010B90400126)

摘　　要：针对船体外板形变曲面实时检测与快速高效三维重建的要求。提出了以空间层次剖分和特征曲率相融合的精简算法,通过k-d树剖分准则将三维点云数据剖分成不同层次空间,层层递归形成树状数据模型,在每个节点空间内,同时分别利用K-邻域计算、曲率估算,获得点云特征曲率信息,设定可调的曲率阈值,依据阈值将同一数据源的点云数据区分为不同曲率大小的区域,运用不同的精简算法,实现保持曲面基本特征的曲面重建。实验结果分析,该算法保证曲面重建的基础上,大大减少了曲面的点云数量,提高了曲面重建效率。For the plate surface deformation of hull in real-time detection and efficient 3D reconstruction requirment, it is proposed space subdivision and feature curvature fusion of streamline algorithm Through the subdivision rules of k-d tree cut three-dimensional point cloud data space it is divided into different levels, the so-called leaf nodes, layer upon layer the recursive form tree data model. In each leaf node space, respectively calculates using the k -neighborhood curvature estimation to get feature curvature information of point cloud, setting the adjustable curvature threshold. According to the threshold of the same data source, point cloud data is divided into the areas of different curvature size and uses different corresponding algorithm to realize surface reconstructionto as keeping the basic features of surface. Analyzes experimental results, the algorithm ensures curved surface reconstruction, greatly reducing the number of point cloud surface, and improving the time of the surface reconsiruction.

关 键 词：船体外板 散乱点云 空间剖分 特征曲率 数据精简 

分 类 号：TH16[机械工程—机械制造及自动化] TP391.41[自动化与计算机技术—计算机应用技术]