Geographic Indexing Mechanisms of Databases for Real-time Visualisation

Geographic visualisation and other real-time applications require efficient data access. An indexing mechanism, adapted to the characteristics of geographic data, can be used to improve retrieval of geographic data. This paper describes an indexing mechanism based on a method called Ellipsoidal Quadtree, EQT. EQT adapts the indexing structure of a database to the shape of the earth. EQT fulfils the so-called equal-area criterion, which means that all index cells can be constructed having equal area. Quadtree sub-division makes retrieval faster, because the EQT adjusts to the varying number of objects, which occurs in different areas. The mechanism is tested using a road database containing all national roads in Sweden. The achieved performance is good. The mechanism has a potential to work on parallel-connected computers.