期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII - Part 3B
页码：155-159
出版社：Copernicus Publications
摘要：LIght Detection And Ranging (LIDAR) data has been recognized as a valuable data source for mapping and 3D modelling of the Earth surface. Classification of LIDAR data for the purpose of extracting ground, vegetation, and buildings is a preliminary step to build 3D models. This paper presents a classification approach of single return LIDAR data that uses area growing technique to extract patches based on neighbourhood height similarity. The extracted patches are classified according to its area into buildings, vegetation, and ground. The classification technique exhibits fast results as it avoids the iterations needed by many classification techniques while maintaining high accuracy level. The presented technique enables simple tuning of parameters because it is directly related to the data specifications. The boundaries of the extracted buildings are then traversed to detect the significant points that help to build the 3D model. The heights of the significant points are computed using the neighbour ground points. Detailed results are presented to show the effectiveness of the proposed approach
关键词：LIDAR; Classification; Building Extraction; 3D Model
