Hybrid transformation for indexing and searching web documents in the cartographic paradigm

摘要

As information becomes available on the World Wide Web in larger quantities and in more disparate formats and media, adequate search engines and portal services providing search and filtering modalities tailored to the needs of the various communities of users become essential components of the global information infrastructure. Global-Atlas is a geographical search engine leveraging the cartographic paradigm for the indexing and searching of World Wide Web documents. Documents are indexed according to their geographical footprint, i.e. the bounding box of the geographical region to which they are related. Documents are searched by interactively drawing bounding boxes on maps. The effectiveness of Global-Atlas depends on the capability to quickly and to accurately index documents and maps. Maps, however, come in a variety of unspecified coordinate systems and projections. In this paper, we evaluate various surface-fitting techniques for the calibration of maps and devise an original hybrid calibration method from the empirical results obtained.