| 基于多源激光点云融合的建筑物BIM建模 |
| |
| 引用本文： | 刘亚坤,刘会云,李永强,赵上斌,杨亚伦.基于多源激光点云融合的建筑物BIM建模[J].地球信息科学,2021,23(5):763-772. |
| |
| 作者姓名： | 刘亚坤 刘会云 李永强 赵上斌 杨亚伦 |
| |
| 作者单位： | 河南理工大学测绘与国土信息工程学院,焦作 454003 |
| |
| 基金项目： | 国家自然科学基金项目(41771491);国家自然科学基金项目(41701597);国家自然科学基金项目(U1810203);中国博士后科学基金项目(2018M642746) |
| |
| 摘 要： | 墙体、窗户等单元构件是建筑物重要组成部分,精细提取其几何参数及位置信息对于完整表达建筑物整体模型具有重要意义。针对单一点云数据源无法获取建筑物单元构件相关参数并完整表达室内外模型重建问题,本文提出一整套融合室内外多源点云数据的BIM模型重建技术。为验证方法的有效性,选取河南理工大学测绘与国土信息工程学院教学楼为实验区域,室内外数据采集时间为2019年5月。在对实验区域机载、车载和地面点云数据进行预处理的基础上,分别选取各点集共轭特征点,以高精度的地面点云为基准,将机载和车载点云融合到地面点云。为提高后期模型重建精度及处理效率,以点云间最小空间距离的方式剔除重叠区域冗余数据。对建筑物进行整体平面与立面剖切,将剖切面在CAD中进行跟踪绘制二维线划图,将二维线划图导入Revit软件中绘制轴网与标高,并利用提取到的墙体几何参数编辑墙体族类型进行BIM模型重建。根据提取到的窗户几何参数统计其类型并编辑窗户族,将其归为有规律性和无规律性两类,有规律性窗户单元找出其重复性规律及位置控制参数,无规律性窗户单元逐个放置,二者结合优化BIM模型。为验证模型重建精度,选取建筑物代表性立面,以人工实测立面边长为参照,将由点云数据提取到的相对应立面边长及模型边长与之对比分析,其误差集中分布在0.0~0.2 m之间,存在0.2 m以上误差,但大部分在0.3 m以下。实验结果证明了该方法的准确性。
|
| 关 键 词： | 多源LiDAR数据 点云融合 冗余数据剔除 剖切图 二维线划图 几何参数提取 单元构件族 BIM模型 |
| 收稿时间： | 2020-07-17 |
Building BIM Modeling based on Multi-source Laser Point Cloud Fusion |
| |
| LIU Yakun,LIU Huiyun,LI Yongqiang,ZHAO Shangbin,YANG Yalun.Building BIM Modeling based on Multi-source Laser Point Cloud Fusion[J].Geo-information Science,2021,23(5):763-772. |
| |
| Authors: | LIU Yakun LIU Huiyun LI Yongqiang ZHAO Shangbin YANG Yalun |
| |
| Institution: | School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454003, China |
| |
| Abstract: | The wall, window, and other unit components are important parts of the building. It is of great significance to extract their geometric parameters and location information to express the overall model of the building. In view of the fact that a single point cloud data source cannot obtain the geometric parameters and location information of building unit components, but can fully and effectively express the problem of indoor and outdoor model reconstruction, this paper proposes a set of BIM model reconstruction technology that integrates indoor and outdoor multi-source point cloud data. In order to verify the effectiveness of this method, the teaching building of the School of Surveying and Land Information Engineering of Henan Polytechnic University was selected as the experimental area, and the data collection time of indoor and outdoor was May 2019. On the basis of preprocessing the airborne, vehicle-borne, and terrestrial laser scanner point cloud data in the experimental area, the conjugate feature points of each point set were selected respectively, and the high-precision terrestrial laser scanner point cloud was taken as the reference to fuse airborne and vehicle-borne point clouds with terrestrial laser scanner point cloud, so as to realize coordinate transformation and reduce the fusion accuracy by iteration. In order to improve the accuracy and processing efficiency of model reconstruction in the later stage, redundant data in overlapping areas were eliminated by means of minimum space distance between point clouds. The overall plane and elevation of the building were dissected, the profile was tracked in CAD to draw two-dimensional line drawings, the two-dimensional line drawings were imported into Revit software to draw axis network and elevation, and the extracted wall geometric parameters were used to edit wall family types for BIM model reconstruction. As for the facade window units, according to the extracted geometric parameters, the types were counted and the window family was edited, and they were classified into regular and irregular. Regular window units were found out by the repeatability law and position control parameters of each type, and irregular window units were placed one by one. The two were combined to optimize the BIM model. In order to verify the accuracy of model reconstruction, the representative facade of the building was selected. Taking the artificial measured facade side length as a reference, the corresponding facade side length and the model side length extracted from the point cloud data were compared and analyzed. The errors were concentrated between 0.0 and 0.2 m, with more than 0.2 m errors, but most of them were below 0.3 m. Experimental results show that the method is accurate. |
| |
| Keywords: | multi-source LiDAR data point cloud integration redundant data removal cutting figure two-dimensional line plot geometric parameter extraction family of unit components BIM model |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《地球信息科学》浏览原始摘要信息 |
| 点击此处可从《地球信息科学》下载免费的PDF全文 |
|
