基于车载激光点云数据的高速公路道路标线提取  

Extraction of expressway Markings Based on Vehicle Mounted Laser Point Cloud Data

作　　者：汤敏 李锦明 黄超 高松涛 吴兴礼 TANG Min;LI Jinmin;HUANG Chao;GAO Songtao;WU Xingli(Guangdong Communication Planning and Design Institute Group Co.,Ltd.,Guangzhou 510507,China)

机构地区：[1]广东省交通规划设计研究院集团股份有限公司,广东广州510507

出　　处：《城市勘测》2025年第2期160-164,共5页Urban Geotechnical Investigation & Surveying

摘　　要：为从高速公路车载激光点云数据中有效提取道路标线,提出结合点云反射强度和道路几何属性的强度特征图像阈值分割法。算法首先使用梯度滤波和点云滤波技术,从初始点云中提取地面点点云;再通过法向量夹角阈值约束的区域生长法获得路面点云,将强度信息投影至工程坐标系所在的XOY平面,得到强度栅格影像;最后使用最大类间方差法进行阈值分割,并联合连通域分析和形态学处理,得到道路标线影像。基于广东某条高速公路的点云数据验证,结果表明该算法具有较好的实用性,在工程生产中具备推广价值。In order to effectively extract road markings from vehicle-mounted laser point cloud data,a threshold segmentation method was proposed,where both the reflection intensity and road geometric attributes were used.The gradient filtering and point cloud filtering techniques were firstly applied to extract the ground point cloud from the initial data.Then,the area growth method constrained by the normal vector angle threshold was used,and the pavement point cloud data obtained.Its intensity was further projected to the XOY plane of the engineering plane coordinate system to form the intensity raster image.The threshold segmentation,connection domain analysis and morphological processing was then applied,the road marking image finally extracted.To evaluate the effectiveness of the method,the point cloud data from Guangdong Expressway were used,showing the practicability.

关 键 词：车载三维激光 激光点云 阈值分割 道路标线提取 

分 类 号：P225.2[天文地球—大地测量学与测量工程]