Improved Classification of Building Infrastructure from Airborne Lidar Data Using Spin Images and Fusion with Ground-Based Lidar

Over the last five to ten years, Airborne Laser Swath Mapping (ALSM) technology, also known simply as Lidar, has become widely available to the remote sensing research community. In that time, many fields of application have been proposed. The majority of the early successful applications were for terrain mapping and surveying, but the high resolution data delivered by this technology has allowed researchers to apply it to problems in forestry, civil engineering, urban planning, and many other fields. In urban planning for example, explicit three-dimensional (3D) building models are an important data product for estimating microwave line-of-sight communications and emergency responder plans. Such models can be extracted from the 3D ALSM point cloud data once the points are properly segmented into building and non-building classes. The work presented here focuses on improved segmentation of building points in the ALSM point cloud using a local subspace mapping known as a spin image. The ALSM data are then fused with ground-based laser scanner data to improve discriminability among complex building architectures, adjacent vegetation, and even rooftop infrastructure.