Map Generalization z Introduction z Concepts xconventional cartography

Map Generalization z. Introduction z. Concepts xconventional cartography xgeographic information systems z. Developments xconceptual models xalgorithms xknowledge representation Image Processing Division

Introduction z. Data presentation xdisplay xcommunication z. Data integration xscale and spatial resolution xdata quality z. Derivation of spatial databases xspatial modeling Image Processing Division

Concepts z. The role of a map is to present a factual statement about geographic reality (Robinson, 1960). z. A map is a data model that intervenes between reality and database (Goodchild, 1992). Image Processing Division

Concepts z. Map generalization is the simplification of observable spatial variation to allow its representation on a map (Goodchild, 1991). z. Map generalization is an informationoriented process intended to universalize the content of a spatial database for what is of interest (Müller, 1991). Image Processing Division

Concepts z. Map generalization: xreduces complexity xretains spatial and attribute accuracy xaccounts for map purpose and scale xprovides more ‘information’ or more efficient communication Image Processing Division

Feature coalescence (Mc. Master and Shea, 1992) Image Processing Division

Feature selection (Monmonier, 1991) Image Processing Division

Complexity reduction (Mc. Master and Shea, 1992) Image Processing Division

Attribute accuracy (Mc. Master and Shea, 1992) Image Processing Division

Map purpose (Mc. Master and Shea, 1992) Image Processing Division

Developments z 1960 to 1975: algorithm development, with emphasis on line simplification. z. Late 1970 s to 1980 s: assessment of algorithm efficiency. z 1990 s: conceptual models; formalization of cartographic knowledge. Image Processing Division

Developments z. Seminal attempts at automation x. Julien Perkal: concept of approximate length of order , where is a real number. x. Waldo Tobler: computer rules for numerical generalization. x. Friedrich Töpfer: amount of information that can be shown per unit area decreases according to geometric progression. Image Processing Division

Conceptual models z. Brassel and Weibel xstructure recognition – measures of relative importance xprocess recognition – definition of generalization process xprocess modeling – compilation of rules xprocess execution – generalization of original database xdata display Image Processing Division

Conceptual models z. Mc. Master and Shea xwhy? – Complexity reduction, maintenance of spatial and attribute accuracy, map purpose and intended audience, retention of clarity xwhen? – Geometric conditions, spatial and holistic measures, transformation control xhow? – Spatial and attribute transformation Image Processing Division

Algorithmic approach z. Overemphasis on line simplification z. Lack of a theory to explain which algorithm is the most appropriate for which object z. Obscure view of what is exploitable z. Necessity to derive methods from semantic and topology rather than from form and size Image Processing Division

Algorithmic approach z. Douglas and Peucker (1973) xredundancy in the number of points of digital lines z. Cromley (1992) xmodification of the Douglas-Peucker algorithm xhierarchical structure to store ranked points z. Li and Openshaw (1992) xconcept of the smallest visible object xhybrid vector/raster implementation Image Processing Division

Algorithmic approach z. Visual comparisons - perception Attneave’s cat (1954) z. Geometric measures xchange in the number of coordinates xchange in angularity xvector displacement xareal displacement Image Processing Division

Knowledge representation z. Knowledge acquisition xconventional KE techniques - communication? xanalysis of text documents xcomparison of map series xmachine learning and neural networks xamplified intelligence Image Processing Division

Knowledge representation y. If expert systems are to be based upon a consensual knowledge of experts, the map generalization realm will not be suited to expert systems technology (Rieger and Coulson, 1993). y. Cooperative knowledge systems should result from joint research in AI, cognitive science, work psychology, and social sciences (Keller, 1995). Image Processing Division

Research agenda y. Objectives of generalization in the digital context y. Test scenarios to push the usefulness of existing tools to their limits y. Cartographic x model-oriented generalizations y. Explicitness of spatial relations for points, lines, and polygons y. Research cooperation between mapping agencies and academia Image Processing Division