ABSTRACT
This chapter focuses on the extraction of three-dimensional (3D) planar primitives. It presents two approaches for detecting 3D roof facets; the first one is based on the analysis of the 3D point cloud while the second one integrates aerial images. The extraction of lines from a Light Detection and Ranging (LiDAR) point cloud is a difficult task since LiDAR points are randomly distributed over surfaces: depending on the point density, building edges are approximately delineated. Integrating aerial images with LiDAR data in a primitive detection process may highly enhance the resulting facets. Automatic mapping of urban areas from aerial images is a challenging task for scientists and surveyors because of the complexity of urban scenes. The chapter shows that searching for planar primitives in a LiDAR point cloud has limitations with regard to a joint use of aerial images and LiDAR data.
