Potential of Airborne Laser Scanning Data for Classification of Wadden Sea Areas

The classification and mapping of habitats in Wadden Sea areas is an important issue of marine monitoring. In the framework of a German research project, we investigate different modern remote sensing data for this task. In this paper, our focus is on the potential of airborne laser scanning data for the classification. Therefore, we use Conditional Random Fields (CRF), a probabilistic supervised classification approach capable of modelling context. We classify the laser scanning point cloud into the three object classes water, mudflat, and mussel bed. For the distinction of different surface types we analyse crucial classification features based on the geometry and the intensity of the backscatter. We then learn typical structures in a training step and combine local descriptors with context information in a CRF framework. We evaluate our approach on a test site of the German part of the Wadden Sea and show classification results of multispectral and SAR data which we intend to combine for a marine monitoring concept.