Using Arcgis Model Builder for Object-based Image Classification of Seagrass Meadows

Seagrass meadows are an important coastal habitat serving as a good indicator of healthy coastal ecosystems. Object-based image classification is a new method of seagrass mapping. The goal of the project was to create a GIS model to classify seagrass in the Puck Bay Natura 2000 habitat protected area (Southern Baltic). The technique was used to analyze the seagrass density and fragmentation. To conduct the project aerial photos of seagrass and ArcGIS Model Builder were employed. Two models were built. The first model proved useful for the segmentation and transformation of the photographs to images containing the objects which were used in the maximum likelihood procedure to produce a classification based on seagrass density. The method – when compared to a pixel based method has a better performance especially for sparse seagrass. The second model was used for landscape division index mapping. Both methods proved to be valuable solutions in seagrass monitoring.