基于激光扫描数据的建筑物自动提取方法研究

Research on Automatic Extraction of Building Based on Laser Scanning Data

指出了随着三维激光扫描硬件设备的不断更新发展,面对更庞大的数据量、精度更高的工程应用,迫切需要具有快速度、高精度的方法来实现扫描数据的建模和计算机的自动分类提取。在已有研究基础上,利用基于球面投影构网和FCM聚类混合的点云分类法,对选取的单站地面激光扫描数据进行处理,并分析了其可行性,主要采用了以下几种处理算法进行了比较分析:基于球面投影的空间三角形构网方法;Fuzzy C-Mean(FCM)模糊聚类方法的点云分类;柱面投影和CYLVIEW浏览器设计;软件开发。

With the continuous update and development of hardware equipment of 3D laser scanners and facing with it sapplication in the project which contains larger data and requires higher accuracy, it urgently needs a method of rapid speed and higher accuracy to be adopted in the modeling of scanning data and computer＇s automatic classification and extraction. Based on the current researches, this article processes the ground laser scanning data of single point and analyzes its feasibility by usingpoint cloudsclassification method which includes the spherical projection constructing-- net and the FCM clustering method. Thearticle selects several processing algorithms to make the comparison and a nalysis, includingspatial triangle constructing--net based on spherical projection, point clouds classificationbased on Fuzzy C-Mean（FCM）, the cylindrical projection and design of CYLVIEWbrowser, and software development.