Automated Extraction of Building Geometry from Mobile Laser Scanning Data Collected in Residential Environments

Spatial data collection in urban environments for the extraction of building inventory information is important for many applications such as urban planning, storm water management, hazard mitigation, vulnerability assessment, and loss estimation to name a few. Creating and updating building inventory databases in large and developing urban environments benefits from efficient data acquisition and data processing techniques. This study leveraged and integrated the advantages of groundbased mobile laser scanning and aerial photography through an automated method to extract building inventory information. The integration of terrestrial and aerial data enables the identification of buildings in the dataset and the extraction of both roof polygons and wall footprints. The presented method was evaluated with real datasets collected from a typical residential area. The area of roof polygons extracted by the automated approach differs by less than 5% when compared to manually calculated values. Also the proposed method allows for accurate extraction of walls captured within the point cloud. INTRODUCTION Three-dimensional representations of urban environments and the extraction of building inventory information are important for many applications such as urban planning, storm water management, disaster management, and emergency response to name a few. Because of the need for up-to-date information in inherently intricate urban scenarios, efficient data acquisition technology and automatic data interpretation techniques have been actively studied and implemented in recent years (Maas and Vosselman, 1999; Süveg and Vosselman, 2004; Brenner, 2005; Pu and Vosselman, 2009; Awwad et al., 2010; Tang et al, 2010). In particular, the segmentation of data into individual features, automatic detection of buildings, and extraction of building ground plans and roof/wall dimensions are major tasks in these research studies. 955 Construction Research Congress 2014 ©ASCE 2014