OSCAR: OpenStreetMap Planet at Your Fingertips via OSm Cell ARrangements

Representing and Reasoning with Context-Sensitive Vague Place Descriptions

“…Therefore, the key to this approach is building a knowledge base that captures possible semantics of natural language phrases (a lexicon). Researchers consider a small set of spatial relations for querying, as discussed in [30], [129], which includes near, inside, and within a specified distance and along with specific cardinal relationships [129]. It has also been demonstrated how topological and metric concepts can evaluate the semantic similarities between various natural language queries using a set of 15 spatial relations.…”

“…The available commercial search engines like Google and Bing Maps have been used by many applications for querying and geo-referencing, while current free/OSM-based search engines are still way behind. Our research uses OSCAR, which uses freely available OSM-data and, due to its substring searches, performs better in cases of the unknown spelling or incomplete names (as discussed in [30]) compared with Google and other search engines [30].The next sections will provide us with the processing pipeline and details about how our prototype works.…”

“…After extracting the information using spatio-ontological reasoning, this thesis's last task focuses on composing the queries and scheduling them for mapping the entities.The chapter tends to focus on the final two components of the processing pipeline described in section 7 of chapter 2.The primary objective is to create queries and algorithms that might strengthen the geo-referencing processes. The queries will be in the form of constraints and based on the type and the name extracted in previous chapters.As a query engine we use oscar 2 [31], [30]. oscar is a new geospatial search engine based on freely available OSM-data.…”

“…oscar is a new geospatial search engine based on freely available OSM-data. The framework can process general query expressions involving unions, intersections, set differences, substring searches, and geographic region constraints on the complete OSM planet data set [30]. The motivation here revolves around creating a natural language interface for oscar, which translates text descriptions into OSM queries by considering simple combinations that can help us with different scenarios present in a place description.…”

“…There are three types of data in the OSM: 'nodes' (with latitude and longitude), polygonal paths or 'ways' (consisting of one or more nodes), and 'relations,' which are compositions thereof (i.e., referencing sets of nodes, ways, or other relations). All three forms may be expanded to include tags called key-value pairs, which allow to name and identify the data [30].…”