Can building footprint extraction from LiDAR be used productively in a topographic mapping context?

Abstract

Light Detection and Ranging (LiDAR) is a quick and economical method for obtaining cloud-point data that can be used in various disciplines and a diversity of applications. LiDAR is a technique that is based on laser technology. The process looks at the two-way travel time of laser beams and measures the time and distance travelled between the laser sensor and the ground (Shan & Sampath, 2005). National Mapping Agencies (NMAs) have traditionally relied on manual methods, such as photogrammetric capture, to collect topographic detail. These methods are laborious, work-intensive, lengthy and hence, costly. In addition because photogrammetric capture methods are often time-consuming, by the time the capture has been carried out, the information source, that is the aerial photography, is out of date (Jenson and Cowen, 1999). Hence NMAs aspire to exploit methods of data capture that are efficient, quick, and cost-effective while producing high quality outputs, which is why the application of LiDAR within NMAs has been increasing. One application that has seen significant advances in the last decade is building footprint extraction (Shirowzhan and Lim, 2013). The buildings layer is a key reference dataset and having up-to-date, current and complete building information is of paramount importance, as can be witnessed with government agencies and the private sectors spending millions each year on aerial photography as a source for collecting building footprint information (Jenson and Cowen, 1999). In the last decade automatic extraction of building footprints from LiDAR data has improved sufficiently to be of an acceptable accuracy for urban planning (Shirowzhan and Lim, 2013).