Designing a Map Legend Ontology For Searching Map Content

Map legends are keys to the understanding of symbols used on maps. Without such legends and the knowledge to interpret them, maps are reduced to mere pictures. While it is possible to semiautomatically extract text, contour lines, major transportation infrastructure, and so forth from maps, more complex aspects and the relation between symbols get lost. Even if these features get accurately extracted, they are mostly used to derive digital representations of (old) analog maps. From an information retrieval perspective, facts such as that a certain map contains transportation features organized in a hierarchy of highways, streets, trails, and so forth, remain hidden and therefore can neither be used by machines nor humans to enable a richer search for map contents. In this work, we formalize a first version of a map legend ontology (MLO) that can be used to semantically annotate and query map contents via their legend in a machine-readable manner using Semantic Web and Linked Data technologies. To demonstrate our results, we introduce three different examples of real-world map legends and show how to represent them by using the map legend ontology.