基于点云分割的建筑物立面结构提取方法  

Building Facade Structure Extraction Method Based on Point Cloud Segmentation

作　　者：江旭 敖建锋 邹永康 Jiang Xu;Ao Jianfeng;Zou Yongkang(School of Architectural and Surveying Mapping Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,Jiangxi,China)

机构地区：[1]江西理工大学土木与测绘工程学院,江西赣州341000

出　　处：《应用激光》2024年第1期66-75,共10页Applied Laser

基　　金：江西省教育厅科学技术研究项目(GJJ2200803)。

摘　　要：建筑物立面结构是描述建筑模型的重要特征之一。为解决现有算法对建筑物立面结构提取时间长、空间投影错误和转换后点云缺失等问题,提出一种基于点云分割的建筑物立面结构提取方法。该方法采用单体化聚类分割提取窗户,通过点云降采样算法稀释主体结构以减少点云密度和复杂度;引入法向量算法实现高密度主体点云到边缘点云的提取;通过改进后的三维线段检测算法遍历边缘点云,实现建筑立面结构特征提取。通过自测点云和Semantic3D公开数据的对比试验,结果表明,所提方法能有效提取建筑立面结构,提高效率并更好地滤除无序线段,为点云立面处理和模型构建提供了一种新的结构处理方法。The building facade structure is one of the essential features to describe the architectural model.To solve the problems of the existing algorithms such as the long time to extract the building facade structure,the spatial projection error,and the missing point cloud after conversion,this paper proposes a building facade structure extraction method based on point cloud segmentation.This method uses single clustering segmentation to extract windows,and dilutes the main structure through a point cloud Downsampling algorithm to reduce the density and complexity of point clouds.Introduce standard vector algorithm to remove high-density main point cloud to edge point cloud.Finally,the improved 3D line segment detection algorithm traverses the edge point cloud to achieve feature extraction of building facade structures.Comparative experiments using self-testing point clouds and Semantic 3D publicly available data demonstrate that the proposed method effectively extracts building facade structures,enhances efficiency,and improves the filtering of disordered line segments.This method offers a new approach for structural processing in point cloud facade processing and model construction.

关 键 词：建筑物立面结构 点云分割 三维线段提取 法向量提取 点云降采样 

分 类 号：P237[天文地球—摄影测量与遥感]