Thematic Mapping from Imagery: an Aspect of Automated Map Generalization

Map generalization, as a means of portraying the complex real world, has to date been confined to the manipulation of map data. With the advent of new data acquisition techniques, particularly remote sensing, data sources for spatial analysis have greatly increased. However, map generalization from image data is a challenging problem. In this paper, the conceptual and technical problems in generalizing cartographic objects from remote sensing imagery are addressed. A two-stage generalization framework is proposed for thematic mapping from imagery. Specific interest is focused on mapping land use from SPOT satellite imagery.