High quality cartographic presentation of spatial information is a basic task and subject matter of geosciences. According to Mackaness et al. (2007), the production of lasting analogous or digital maps with a rich variety of content is nowadays replaced by production of updated, easily accessible and specialized thematic maps. Therefore, proper and automated generalization of spatial information through a variety of applications or services plays a significant role. The benefit of automatic generalization can be seen in creation of cartographic representations in different scales from the same spatial database, which is a fundamental requirement of the majority of National Mapping Agencies (Regnauld et al, 2007).

Cartographic generalization is a complex process which requires cartographer's interdisciplinary approach, respecting various attributes of spatial data and adoption of multiple geometric rules. Because of the complexity and multiple factors influencing the choice of objects to be displayed and the way they are represented in the generalized data set, procedures of automatic generalization process are complicated to implement in the digital environment. Development of multi-agent systems is nowadays one of the main ways of designing systems for automatic generalization and as such the most studied method of implementation of automated cartographic procedures in the digital environment.

This contribution presents concepts and the development of multi-agent systems for use in automatic generalization. It also presents a system for automated generalization of spatial data created using the Safe Software FME Desktop software package primarily designed for conversion, transformation, and integration of spatial data. The created system is based on constraints – the required conditions that elements of cartographic representation derived from the original cartographic spatial database have to meet. The system enables generalization of polygonal objects (buildings) and line objects (roads) according to the requirements for displaying on a map at the scale of 1:50 000. The presentation presents advantages and shortcomings of this approach to cartographic generalization, the quality of the generalization is analysed and possible improvements to the system are listed.

References

Mackaness, W. A.; Ruas, A.; Sarjakoski, L.T. (editors) (2007): Generalisation of Geographic Information: Cartographic Modelling and Applications. International Cartographic Association, Elsevier, Amsterdam.

Regnauld, N.; Féchir, A.; Lecordix, F.; Rejkjær, D. (2007): NMA's Collaboration on Generalisation the Magnet Consortium. XXIII International Cartographic Conference, Moscow, Russia.

http://icaci.org/documents/ICC_proceedings/ICC2007/html/Proceedings.htm (11.5.2011).