Urban land cover and land use extraction from very high resolution remote sensing imagery

Urban land cover and land use information is essential for many urban-related applications. It is known that remote sensing images have the potential of extracting urban land cover and land use information and monitoring its change at local, regional, and national scales. In particular at local scale, the growing availability of Very High Resolution (VHR) remote sensing images, e.g., from QuickBird, GeoEye, WorldView and Pleiades sensors, has caused a considerable increase in both scientific and application fields associated with urban land cover and land use extraction. By definition, land cover refers to the physical properties of the each surface, whereas land use refers to the corresponding functional aspects, i.e. how the land cover is used by human beings.
This paper presents a novel method for urban land cover extraction from VHR images, and models the link between land cover and land use for urban land use extraction. For urban land cover extraction, from the methodological point of view, we focus on building roofs extraction from single VHR imagery by making use of the directional relationship between a building roof and its shadow. For modelling the link, we provide a novel way to statistically quantify the spatial arrangement of land cover elements for characterizing urban land use. Then, the urban land use classification is conducted. We applied our proposed method to a subset of Pleiades image at an urban area of Wuhan, China. From our experiments, we conclude that our proposed method can provide an effective means for urban land cover and land use extraction. In addition, the challenges and future work associated with urban land cover and land use extraction are discussed.