Robust Approach towards an Automated Detection of Built-up Areas from High Resolution Radar Imagery

The actual process of rapid urbanisation is associated with various ecological, social and economic changes in both, the urban area and the surrounding natural environment. In order to keep up with the effects and impacts of this development, effective urban and regional planning requires accurate and up-to-date information on the urban dynamics. Recent studies have demonstrated the applicability of high resolution optical satellite data for the acquisition of spatial and socio-economic information. In contrast, radar imagery has hardly been employed for these purposes so far. The future TerraSAR-X will provide radar data with a ground resolution comparable to existing high resolution optical satellite sensors. Thus, it will afford detailed urban analysis based on spaceborne radar imagery for the first time. The detection concept presented here serves as a preliminary investigation of the potential use of TerraSAR-X data in the context of urban applications. It introduces a robust approach towards an automated detection of built-up areas using data acquired by the Experimental Synthetic Aperture Radar (E-SAR) system of the German Aerospace Center (DLR). For that purpose different data sets of single-polarised Xband imagery are analysed in an object-oriented classification. A robust object-oriented analysis strongly depends on accurate and reliable image segmentation. Thus, a classification-based optimisation procedure to stabilise and improve the initial image segmentation step is introduced. Subsequently the identification of built-up areas is performed on the basis of three image segmentation levels in different spatial scales. Here, contextual and textural features along with shape-related and hierarchical characteristics play a major role. Finally, the transferability and robustness of the presented approach is illustrated by applying the developed classification scheme to E-SAR data of three complete flight tracks.