Robust Mapping of Urban Structure Types Across Three German Cities

Urban structure types (USTs) are a concept in urban ecology to divide cities into units of homogenous environmental conditions. They provide a useful means to conceive efficient strategies for sustainable urban development. Cutting edge remote sensing data and methods allow for the automation of the UST classification process. This paper aims at demonstrating the robustness of a recently developed approach for area-wide mapping of 13 USTs. To this end, the experimental setup of this work includes three major cities in Germany, different sets of high-resolution multi-source data as well as statistical measures to enable the assessment of methodological accuracy and transferability. The results of this study emphasize the suitability of the presented approach with regard to classification accuracy, workflow automation, and operational readiness for future UST mapping and monitoring tasks.