Integration of Gis and Object-based Image Analysis to Model and Visualize Landscapes

Although a number of digital terrain analysis techniques were developed since the early 90ies terrain information is not utilized enough in landscape research even if topography is a key variable in a wide range of environmental processes. One explanation of this situation is the two-dimensionality of the main paradigm in landscape ecology, namely the patch-corridor-matrix model. Calculations of landscape metrics based on such definitions of landscape elements can lead to biased or misleading results if the topographic situation plays an important ecological role. Landscape spatial indices (e. g. mean patch size, fragmentation, diversity etc.) are strongly influenced by landforms which act as natural breaks in landscape patterns (steep slopes), or conversely, as spatial connectors (valleys). To account for these relationships, an image segmentation-based methodology of landform classification is proposed with the objective to integrate it into the definition of landscape units.