基于结构单元探测与修复的车载LiDAR数据建筑物立面模型三维重建研究

Due to the massive number points of mobile LiDAR data, the existence of complicated landscapes and complex façade structures, the limitations of data collection, and the occlusion and disturbance by objects on the street (e.g., trees, vehicles), the reconstruction of a high-quality building façade model with a high level of automation presents a challenging but very worthwhile research topic. In response to these issues, this project aims to reconstruct 3D building façade models with high integrity and high quality, plans to propose a technical strategy, as “building points detection –building points unfolding – repeated structures detection – building points restoration – building model construction”. The project put forward some measures to break through the "bottleneck" of such technical difficulties. First, a point cloud detection method is proposed to rapidly and efficiently extract the building façade points, and preliminary split the complex building structures. Second, an unfolding method whereby a building façade is flattened onto a 2D plane is introduced, to reduce the difficulty of building reconstruction for multiple façades. Third, a two-direction detection strategy, with façade structure detection performed in the vertical and horizontal directions successively, is used to determine repeated patterns. Four, a restoration method is developed to consolidate the imperfect data according to the similarity of façade structures. Finally, the high integrity 3D building façade models are reconstructed by using the restored façade points and the repeated façade structures. The method makes use of repeated structures in building’s façades, create new ideas and new methods for the reconstruction of 3D building façade models from mobile LiDAR data, and provide effective solution for the typical applications, such as the geographic spatial information infrastructure of "Smart City".

由于车载LiDAR点云数据量大、完整性低、噪声信息多、场景目标复杂等固有不足，以及建筑物立面层次结构复杂、细节特征丰富等问题，导致LiDAR数据建筑物立面模型三维重建存在数据处理难、点云修复难、模型重建难等关键技术问题。本项目以车载LiDAR数据建筑物立面模型精细化重建为研究目标，拟提出“建筑物点云提取－立面点云展开－结构单元探测－立面点云修复－三维模型重建”技术架构，突破建筑物立面点云的微观细节探测与三维模型重建的技术瓶颈，探索基于车载LiDAR数据的建筑物结构单元探测技术，深入挖掘建筑物立面局部结构的重复性或自相似性特征，降低点云处理与三维建模的复杂性，提升建筑物立面模型的完整精细程度，为大区域城市建筑物目标准确感知与模型精细表达等提供新思路与新方法，为智慧城市中的城市空间信息基础设施建设等典型应用提供有效的解决途径。

项目围绕“数据集成—特征提取—三维建模”的研究思路，针对LiDAR点云数据创新提出了一系列新型技术与方法。首先，提出三种多源LiDAR数据配准方法，包括基于路网的多层次配准方法、基于公共地面点集的配准方法、基于可移动角点的配准方法，集成“俯视+侧视”多视角下LiDAR数据的互补优势，以提高后续特征地物提取与三维模型重建的完整性、精确性和细节性。其次，提出两种特征地物提取与分割方法，包括一种基于虚拟种子点的机载LiDAR点云提取方法和一种侧视LiDAR数据多层次单木点云分割方法，实现了大区域城市环境下地面与树木等特征地物的微观细节探测。最后，形成一套建筑物立面模型重建方案，包括提出一种基于体元组结构与空心率分析的建筑物点云提取方法和一种基于相似结构探测的建筑物立面模型重建方法，实现高完整性的建筑物点云提取、结构单元修复和三维模型重建，实现了项目立项时确立的研究目标。