全局线形特征与局部结构特征约束的隧道激光雷达点云逆向参数化建模方法  

Reverse parametric modeling of tunnel from LiDAR point clouds constrained by global linear features and local structural features

作　　者：张鹏阳 丁雨淋[1] 郝蕊 胡翰[1] 范文娜 程智博 孙安培 朱庆[1] ZHANG Pengyang;DING Yulin;HAO Rui;HU Han;FAN Wenna;CHENG Zhibo;SUN Anpei;ZHU Qing(Faculty of Geosciences and Engineering,Southwest Jiaotong University,Chengdu 611752,China;China Academy of Railway Sciences Corporation Limited,Beijing 100081,China)

机构地区：[1]西南交通大学地球科学与工程学院,成都611752 [2]中国铁道科学院集团有限公司,北京100081

出　　处：《测绘工程》2025年第2期24-31,共8页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(42371436);中国铁道科学院基金资助项目(2022YJ057)。

摘　　要：隧道三维模型是隧道智能运维与精细化管控的重要基础,隧道结构复杂、构件多样,洞内空间狭小且相互遮挡,加之激光雷达点云细节缺失、几何特征不明显,导致逆向精准提取隧道规则结构参数困难。为此,提出了全局线形特征与局部结构特征约束的隧道激光雷达点云逆向参数化建模方法,采用多视角投影及迭代最优估计提取隧道线形特征,构建仰拱、底板等典型基元结构的参数化模版,以全局线形特征约束提取横截面点云,以局部模版几何结构自适应匹配断面点云提取隧道结构参数,基于既有BIM软件参数化重建引擎实现隧道逆向参数化建模。利用典型隧道激光点云数据验证了文中方法的有效性,结果表明,文中方法能高效、完整提取线性参数和结构参数,半径拟合标准差为3 mm,隧道自动化建模可达mm级。The 3D model of a tunnel is a crucial foundation for intelligent operation,maintenance,and refined control of tunnels.Tunnels have complex structures and diverse components,with narrow internal spaces prone to mutual obstruction.Additionally,the lack of detail in LiDAR point clouds and inconspicuous geometric features make it challenging to accurately reverse-extract the structural parameters of tunnels.To address these issues,reverse parametric modeling of tunnel from LiDAR point clouds constrained by global linear features and local structural features is proposed.This method employs multi-view projection and iterative optimal estimation to extract the linear features of the tunnel.It constructs parametric templates for typical primitive structures such as upward arches and base plates.Cross-sectional point clouds are extracted using global linear feature constraints and further segmented using the adaptive matching of local template geometric structures to extract structural parameters.The reverse parametric modeling of the tunnel is realized through the parametric reconstruction engine of existing BIM software.The method's validity is verified using LiDAR point cloud data from typical tunnels.The results demonstrate that the proposed method can efficiently and comprehensively extract linear and structural parameters,achieving a standard deviation of 3 mm in radius fitting,enabling millimeter-level precision in automated tunnel modeling.

关 键 词：激光雷达点云 逆向建模 参数化 隧道基元 

分 类 号：P231[天文地球—摄影测量与遥感] TP751[天文地球—测绘科学与技术]