面向3D视觉引导的点云分割  

Point Cloud Cegmentation for 3D Visual Guidance

作　　者：周洪志 杨海波 贾军营 卢鑫 李子琦 ZHOU Hong-Zhi;YANG Hai-Bo;JIA Jun-Ying;LU Xin;LI Zi-Qi(School of Information Science and Engineering,Shenyang University of Technology,Shenyang 110870,China;Shenyang Fengchi Software Co.Ltd.,Shenyang 110117,China)

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870 [2]沈阳风驰软件有限公司,沈阳110117

出　　处：《计算机系统应用》2024年第10期236-244,共9页Computer Systems & Applications

基　　金：辽宁省自然科学基金(2022-MS-438);辽宁省“揭榜挂帅”重点科技攻关项目(2022JH1/10800085);辽宁省教育厅基本科研项目(LJKFZ20220184)。

摘　　要：点云分割是三维视觉引导和场景理解中的关键步骤,点云分割的质量直接影响三维测量或成像的质量.为提高分割精度、解决边界越界问题,本文提出了一种面向3D视觉引导的点云分割算法,该算法根据点云的空间位置、曲率和法向量信息,生成初始超体素数据,并提取边界点;通过计算边界点与邻域超体素的相似性度量,进行边界细化,即重新分配边界点优化超体素;最后基于区域生长获得候选片段并根据其凹凸性进行合并,得到对象级分割结果.经过可视化和定量比较表明,该算法有效解决了边界越界问题,能对复杂的点云模型准确分割,分割结果准确率为89.04%,召回率为87.38%.Point cloud segmentation is a crucial step in 3D visual guidance and scene understanding,whose quality directly affects the quality of 3D measurement or imaging.To improve the segmentation accuracy and solve the out-ofbounds problem,this study proposes a point cloud segmentation algorithm for 3D vision guidance.This algorithm generates initial supervoxel data and extracts boundary points based on the spatial position,curvature and normal vectors of the point cloud.Boundary refinement is then performed,which refers to the redistribution of boundary points to optimize supervoxels,by calculating the similarity measure between boundary points and neighboring supervoxels.Ultimately,candidate fragments are obtained based on region growing and merged according to their concavity and convexity to achieve object-level segmentation.Visualization and quantitative comparison show that this algorithm effectively solves the out-of-bounds problem and accurately segment complex point cloud models.The segmentation accuracy is 89.04%and the recall rate is 87.38%.

关 键 词：点云分割 超体素 边界细化 凹凸性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]