Residential Building Reconstruction Based on the Data Fusion of Sparse LiDAR Data and Satellite Imagery

Abstract

Reconstruction of residential building models in an urban environment is a challenging task yet has many applications such as urban planning, simulation of disaster scenarios, cartography, wireless network planning, line-of-sight analysis, virtual tours, and many others. This paper presents a novel method for residential building reconstruction and modeling in urban areas using airborne light detection and ranging (LiDAR) data and satellite imagery. The main contribution is the automatic isolation of building roofs and roof reconstruction based on the fusion of LiDAR data and satellite imagery. By using cue lines which are generated from satellite imagery to separate buildings from other objects (including other buildings), we are able to automatically identify individual buildings from residential clutter and re-create a virtual representation with improved accuracy and reasonable computation time. We applied the method to urban sites in the city of New Orleans and demonstrated that it identified building measurements successfully and rendered 3D models effectively. Our experiments show that our method can successfully reconstruct small buildings with relatively sparse LiDAR sampling and in the presence of noise.