Investigating Urban Railway Corridors with Geometric High Resolution Satellite Data

Accurate analyses concerning the ecological value of urban railway sites are required to ensure a sustainable development of inner urban open space. One problem in this context is the great heterogeneity of such areas. New opportunities to investigate heterogeneous urban areas arose with the advent of geometric very high resolution (VHR) satellite data. A bundle of panchromatic and multispectral QuickBird data was employed to explore the potential of such sensors for urban ecological monitoring and assessment exemplified for a railway area in the Southeast of Berlin. An object-oriented, hierarchical classification approach was employed to integrate shape and texture parameters into the analysis. Seven land use classes based on three hierarchical object levels were distinguished. A validation based on field mapping and aerial photography revealed an overall accuracy of about 85%. The classification result allows to answer various ecologically determined questions and to describe the site’s inherent ecological value. Imperviousness was calculated based on the material specific degree of sealing and different vegetation succession states were differentiated. Moreover, the separation of vegetation on tracks from other vegetation allowed deriving information about the recent utilization of railway tracks, which is not available otherwise. It is concluded that the geometric resolution of QuickBird VHR data is suited to separate relevant surface components of derelict urban railway sites at a local scale.