Geodetic Distance Queries on R-Trees for Indexing Geographic Data

Abstract

AbstractGeographic data have become abundantly available in the recent years due to the widespread deployment of GPS devices for example in mobile phones. At the same time, the data covered are no longer restricted to the local area of a single application, but often span the whole world. However, we do still use very rough approximations when indexing these data, which are usually stored and indexed using an equirectangular projection. When distances are measured using Euclidean distance in this projection, a non-neglibile error may occur. Databases are lacking good support for accelerated nearest neighbor queries and range queries in such datasets for the much more appropriate geodetic (great-circle) distance. In this article, we will show two approaches how a widely known spatial index structure – the R-tree – can be easily used for nearest neighbor queries with the geodetic distance, with no changes to the actual index structure. This allows existing database indexes immediately to be used with low distortion and highly efficient nearest neighbor queries and radius queries as well as window queries.KeywordsIndex StructureRange QueryQuery PointGeodetic DataNeighbor QueryThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.