最小二乘角度约束的激光点云标线边界特征提取方法  

Extraction Method for Laser Point Cloud Marking Boundary Features with Least Squares Angle Constraint

作　　者：张晓鹏 刘如飞[1] 谢永强 李明 马新江 ZHANG Xiaopeng;LIU Rufei;XIE Yongqiang;LI Ming;MA Xinjiang(College of Geodesy and Geomatics,Shandong University of Science and Technology,Qingdao,Shandong 266590,China;Qingdao Digital World Information Technology Co.Ltd.,Qingdao,Shandong 266590,China;Department of Survey Science and Engineering,Hohai University,Nanjing 210000,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛266590 [2]青岛数字世界信息科技有限公司,山东青岛266590 [3]河海大学测绘科学与工程系,南京210000

出　　处：《遥感信息》2025年第1期165-172,共8页Remote Sensing Information

基　　金：国家自然科学基金(42001414);“菁英计划”科研支持经费(0104060541613)。

摘　　要：针对道路激光标线点云边界特征的过欠分割问题,提出了一种基于角度约束的最小二乘标线边界角点提取方法,实现对标线边界特征的准确提取,从而构成标线轮廓。首先,对原始点云采用OTSU阈值分割将导向标线点云与路面点云分离,并通过基于KD-tree的区域生长优化阈值分割的结果,同时去除离群点,获得标线点云;随后,利用Alpha shapes提取标线轮廓,通过RANSAC算法对特征边界进行分割;最后,通过最小二乘角度约束的平差模型,以获得边界最优的角点,从而构成规则的标线。文章重点讨论了直线-直线边界的边界特征相交以及直线-圆弧边界特征相交两种情况。实验结果表明,与常用的最小二乘拟合相交获取角点的方法相比,该算法在提高效率的同时,能够获得更高边界轮廓线的精度。Aiming at the problem of over and under segmentation of the boundary features of the laser marking point cloud,this paper proposes a method of extracting corner points of the marking line boundary based on angle constraints to achieve accurate extraction of the boundary features of the marking line,so as to form the marking line contour.Firstly,the original point cloud is separated from the road point cloud by OTSU threshold segmentation,and the threshold segmentation results are optimized based on KD-tree region growth,and the outliers are removed to obtain the marking point cloud.Then,Alpha shapes are used to extract the line contour,and the feature boundary is segmented by RANSAC algorithm.Finally,the optimal corner of the boundary is obtained by means of the adjustment model with least square angle constraint,and the regular line is formed.In this paper,two cases of boundary feature intersection of straight-line boundary and straight-arc boundary feature intersection are discussed.The experimental results show that compared with the common method of least square fitting intersection to obtain corner points,the proposed algorithm can obtain higher precision of boundary contour while improving efficiency.

关 键 词：标线点云分割 Alpha shapes边界提取 RANSAC边界分割 最小二乘角度约束 

分 类 号：TN249[电子电信—物理电子学]