TERRASAR-X AND RAPIDEYE DATA FOR THE PARAMETERISATION OF RELATIONAL CHARACTERISTICS OF URBAN ATKIS DLM OBJECTS

Abstract. This work presents a multi-sensor data analysis concept for the parameterisation of urban landuse in comparison to ATKIS DLM reference objects (digital landscape model). An object based top-down approach is implemented and the potential of multisensor data for a primary urban landcover object classification is assessed. Urban landuse structure is developed based on relational features applied to land cover objects and compared to an aggregated DLM class legend. For a better imperviousness description an advanced imperviousness measure – the build-up impervious intensity ratio is developed that takes building height and derived floor area (based on LiDAR data) and the amount of vegetation within a search radius to every building into account. The developed concept is parameterized with test areas in Rostock and transferability is investigated with data coverage in Cologne (Germany). The work is linked to the urban work package of the DLR funded ENVILAND-2 project that aims to develop operational concepts for the use of TerraSAR-X and Rapideye data in urban mapping scenarios.