Detection of building outlines based on the fusion of SAR and optical features

This paper deals with the automatic extraction of building outlines using a pair of optical and synthetic aperture radar (SAR) images. The aim is to define areas of interest for building height reconstruction in radargrammetric or interferometric applications. Since high resolution optical satellite images are now easily available, such methods merging SAR and optical information could be useful to improve 3D SAR reconstruction (the optical image giving only information on the scene organization). Both SAR and optical data bring complementary information about the building presence and shape. The proposed method is divided into two main steps: first, extraction of partial potential building footprints on the SAR image, and then shape detection on the optical one using the previously extracted primitives (lines). Two methods of shape detection have been developed, the simplest one finding the "best" rectangular shape and the second one searching for a more complicated shape in case of failure of the first one. Results for an industrial area acquired with two incidence angles for the SAR image are presented and analyzed.