GIS-embedded remote-sensing image analysis

In this paper an integration scheme is proposed and applied for the updating of land-cover databases by means of remote-sensing imagery, based on the information content of available imagery. After a preliminary analysis of the neighboring polygons (i.e., background) around a polygon to update (i.e., object), the best feature is selected out of a set of more than 30 features, based on its ability to separate object from background. This best feature is used in a successive image labeling step, after which the labeled image is transformed into a new polygon. This process is repeated for all the polygons that must be updated. In this fashion, only a local section of the image is interpreted with a given feature vector, making the analysis more robust to atmospheric and soil conditions than classical approaches using a fixed feature or feature vector for the entire scene.