Validation of the European Urban Atlas & Automatic Classification of Urban Structures utilizing Cartosat-1 nDSM Data

Ongoing urbanization and population growth affect physical, ecological and socio-economic processes within and beyond urban areas. In this context, the physical spatial structure of cities serves as a basis for understanding urban areas in their entirety in order to achieve a sustainable urban growth. Remote sensing is a cost-effective data source to derive information on the physical spatial structure of urban areas in large areal and temporal coverage. The European Urban Atlas dataset is generated mainly on the basis of remotely sensed data and provides information on landuse/landcover of cities with more than 100,000 inhabitants in Europe. However, as proved within this study, the classification scheme does only partially hold distinct information on building structures and thus, as a matter of fact, does not contain holistic information on the morphology of urban areas. Within this study, a methodology towards the delineation of urban structures utilizing Cartosat-1 nDSM data and Urban Atlas building block information was developed. The nDSM data proved to be substantial in terms of building height classification but fails to add information on the building density. The methodology was applied in another city context achieving constant results and to an artificial reference unit of square objects achieving worse but still satisfactory results. The latter overcomes spatial limitations and allows for deriving information on urban structures independently from additional spatial data sources, on continental and even global scale. A cross-city structural analysis revealed differences and analogies of the urban morphology of Paris and London depending on the distance to the respective city center