New Advances in Automated Urban Modelling from Airborne Laser Scanning Data

Traditionally, urban models in many applications such as urban planning, disaster management, and computer games only require visual accuracy. However, more recently, updating urban infrastructure combined with the rise of mega-cities (i.e. those with populations over ten million) has motivated researchers and users to utilize city-scale models for engineering purposes (e.g. tracking pollution monitoring, optimizing solar panel placement), which necessitates high geometric accuracy. Currently, a major bottleneck lies in the cost of generating accurate, geo-spatially referenced models. This paper presents the evolution of some of the efforts to automatically produce such models. Specifically, recent advances in airborne laser scanning can rapidly acquire accurate, spatial data for large geographic areas in hours, but due to the size of the data sets, coupled with difficulties of capturing and portraying complex structures, many post-processing issues have only recently been addressed to a level sufficient to begin to facilitate automation, especially of building surface reconstruction. Automation is a critical step for further processing and utilization of airborne laser scanned data for engineering-based, urban modeling. This paper presents recent development of the methods for building detection and extraction, with an emphasis on patents and other contributions related to automated processing of airborne laser scanning data.Deposited by bulk impor