Integration of multi-sensor data for landscape modeling using a region-based approach

The increased availability of multi-sensor data, and elevation information in particular, leads to the need of advanced processing methods. In the context of landscape modeling tasks, we concentrate on one central component, the extraction of terrain surface from a Digital Surface Model (DSM). In contrast to conventional mathematical grey value morphology approaches (filtering methods) or to stochastical procedures, we propose an alternative methodology for this task by applying a region-based and multi-scale approach. It consists of segmentation and follow-up fuzzy logic classification based on several features derived from elevation and multi-spectral image data. The satisfying results obtained with a multi-sensor as well as with other datasets show the applicability of the approach.