地铁隧道海量点云管理与系统设计研究

发布时间：2018-08-30 12:09

【摘要】：传统的地铁隧道沉降监测主要通过人工在隧道中每隔一定距离布设断面点,使用全站仪进行观测,该方法监测点数量有限,并且测量精度易受环境因素的影响。而病害检测则是采用人工检测的方法,这种方式不仅速度慢,而且由于人眼观察距离有限,极易出现遗漏的问题。三维激光扫描技术是近几年快速发展的非接触测量技术,将三维激光扫描技术应用于隧道的沉降监测和病害检测,不仅能够提高工作效率,还能够获取高密度的隧道表面点云数据,方便隧道的建模。海量的隧道点云数据给计算机软硬件带来了巨大的考验,计算机的内存和CPU处理能力有限,不能同时对所有点进行处理,因此需要合理的索引结构对隧道点云进行管理。目前国内外有很多学者对海量点云的管理与可视化的理论进行了相关研究,提出了不同的点云管理模型,如四叉树模型、八叉树模型、R树模型等等。每一种点云管理模型的适用情况是有限的,并不能适合所有不同空间分布的点云。本文采用隧道扫描小车获取点云的方式,介绍了隧道扫描小车的基本硬件组成,以及使用扫描小车获取点云的方法。由于扫描的动态性,本文还介绍了点云的配准方法。为了管理获取的海量隧道点云,结合隧道点云的空间分布特点,本文提出了适用于管理隧道点云的混合索引模型以及高效的点云数据存储方式。为了实现流畅的点云可视化效果,本文提出了基于OSG的内外存调度的隧道海量点云的可视化方法,通过使用OSG提供的分页数据库技术,结合建立的隧道点云索引结构,实现了流畅的可视化效果。基于隧道点云的管理与可视化理论,本文设计并实现了隧道点云处理平台,该平台集隧道点云的管理、可视化、交互于一体,并将隧道的横断面点云提取的功能集成到该平台当中,方便时实际的工程应用。最后,本文选取武汉市某地铁线路一段隧道作为实验区,使用扫描小车对隧道进行了扫描实验,将获取点云使用开发的点云处理平台进行了相关处理,处理的效果显示该点云管理与可视化理论满足隧道海量点云处理的需求。
[Abstract]:The traditional subsidence monitoring of metro tunnels mainly uses total station to observe the subsidence of the tunnel by setting section points at certain distances manually. The number of monitoring points is limited and the measuring accuracy is easily affected by environmental factors. The method of disease detection is manual detection, which is not only slow, but also due to the human eye. Three-dimensional laser scanning technology is a non-contact measurement technology developed rapidly in recent years. The application of three-dimensional laser scanning technology to tunnel settlement monitoring and disease detection can not only improve work efficiency, but also obtain high-density tunnel surface point cloud data, which is convenient for tunnel modeling. Massive tunnel point cloud data has brought tremendous challenges to computer software and hardware. The computer memory and CPU processing capacity are limited, and all points can not be processed at the same time. Therefore, a reasonable index structure is needed to manage tunnel point cloud. Different point cloud management models are proposed, such as quadtree model, octree model, R-tree model and so on. Each point cloud management model is limited and can not be suitable for all point clouds with different spatial distribution. This paper also introduces the registration method of point clouds due to the dynamic nature of scanning. In order to manage the massive tunnel point clouds acquired, a hybrid index model for managing tunnel point clouds and efficient point cloud data storage are proposed in this paper. In order to achieve a smooth visualization of point clouds, this paper proposes a visualization method of massive tunnel point clouds based on OSG memory scheduling. By using the paging database technology provided by OSG and the index structure of tunnel point clouds, the visualization effect is realized smoothly. This paper designs and implements a tunnel point cloud processing platform, which integrates the management, visualization and interaction of the tunnel point cloud, and integrates the function of extracting the cross-section point cloud of the tunnel into the platform for convenient and practical engineering application. Finally, this paper selects a tunnel of a subway line in Wuhan as the experimental area, using the scanning trolley. Scanning experiments were carried out on the tunnel, and the point cloud processing platform was used to acquire the point cloud. The results show that the point cloud management and visualization theory can meet the needs of massive point cloud processing in the tunnel.
【学位授予单位】：武汉大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U456.3;U231

【相似文献】