A framework to formalize multi-scales land-use generalization in the ScaleMaster 2.0
1Girres, J.; 2Touya, G.
1UMR GRED – UNIVERSITÉ PAUL VALÉRY MONTPELLIER Email: firstname.lastname@example.org
2IGN - COGIT Email: guillaume.touya@ign
Multi-representation databases (MRDB) allow representing the same real world entities in different cartographic databases with their own level of detail. It enables navigating from a scale to another using multi-scales geoportals. To facilitate smooth transitions (i.e. cartographic continuum) between representations, cartographic generalization allows deriving intermediate representations, based on the most detailed database. In this context, the ScaleMaster2.0 (Touya & Girres, 2013) was developed as a multi-scales generalization system, allowing the derivation of intermediate levels of detail by formalizing generalization rules from a MRDB. This abstract describes how the generalization of land-use data can be performed within the ScaleMaster2.0 model. Land-use coverages constitute a partition of the real world, where each land-use class can be handled differently in a generalization process. For instance, a small polygon can be enlarged if it represents an isolated urban area, but can be deleted if it represents a piece of land. Thus, land-use generalization processes require specific rules for each land-use class, according to the geographical context of the objects. Several land-use generalization models have been proposed in the literature, as for instance models guided by the geometry (van Oostrom, 1995), by the data schema (where two object classes are merged at a smaller scale), or using mixed approaches (Haunert & Wolff 2010). In this context, this on-going research proposes a general framework to formalize generalization processes for land-use coverages in the ScaleMaster2.0 model, by combining operations guided by the geometry and the data schema, and also taking into account computational limitations involved by the extent of the land-use coverage to generalize. The developed method formalizes for each land-use class two parameters are defined: a merging option to another class and a minimal area threshold. These parameters need to be defined for each scale range of the ScaleMaster if generalization has to be performed at any scale. On a functional point of view, the developed land-use generalization process follows the steps: 1. partition the land-use coverage in regular squares (if needed, to optimize computation), 2. merge the classes, 3. Convert the land-use coverage into a topological map, 4. Eliminate the small objects, 5. Apply line simplification algorithms. The developed land-use generalization method was experimented on two different case studies: a land-use coverage in the area of Dakar (Senegal) and a map of sedimentary funds off the coast of Brittany (France). The two experiments show interesting results to derive intermediate land-use coverages from a MRDB, with acceptable time computation by the use of partitioning options of the land-use coverages. Nevertheless, improvements should be provided concerning the consistency between neighboring partitions of the generalized land-use coverages, as proposed by von Oostrom (1995) and with lower level of details, as evoked by Girres & Touya (2014). References Girres JF, Touya G (2014) Cartographic Generalisation Aware of Multiple Repre-sentations, Proceedings of the 8th International Conference on Geographic In-formation Science , Vienna, Austria Haunert JH, Wolff A (2010) Area aggregation in map generalisation by mixed-integer programming. International Journal of Geographical Information Science 24(12):1871–1897 Touya G, Girres JF (2013) ScaleMaster 2.0: a ScaleMaster Extension to Monitor Automatic Multi-Scales Generalizations. Cartography and Geographic Infor-mation Science 40(3):192–200 van Oosterom P (1995) The GAP-tree, an approach to 'on the fly' map generaliza-tion of an area partitioning. In: Müller JC, Lagrange JP, Weibel R(Eds.), GIS and Generalization: Methodology and Practise, pp. 120–132. London: Taylor & Francis.
generalization; land-use; scalemaster