ABSTRACT
This chapter describes the algorithms that derive building models and examines the results by applying the proposed framework to the sample data set. It presents a framework involving a series of algorithms for the extraction of simple and sophisticated building models from Light Detection and Ranging (LiDAR) measurements. The framework consists of three major steps. First, the nonground and ground measurements are separated. Second, building measurements are identified by region growing using a local plane-fitting technique. Third, simple building models are derived and adjusted based on estimated dominant directions, and sophisticated building models are derived and refined based on the 2D topology of roof facets. Airborne LiDAR systems generate voluminous and irregularly spaced three dimensional point measurements of objects, including ground, building, trees, and cars scanned by the laser beneath the aircraft. The sheer volumes of point data require dedicated algorithms for automated building reconstruction.
