Hierarchical data structures for geographical objects
Most of the current geographical information systems (GIS) are built for a single scale representation. They are developed on systems which do not usually capture the semantics and the inherent structure of geographical objects. The models are usually based on restricted data types, and the relationships among the objects are generally unnaturally represented. This thesis introduces a hierarchical geographical structure called IMAS (Intelligent MAp Structure). It is based on the frame structure of expert systems and provides a unifying framework for multiple scale representation of geographical objects. A prototype of this structure has been implemented and tested with EMR map data at scales of 1:2,000,000 and 1:7,500,000. For this data, IMAS required approximately twice as much space as the standard EMR format. Substantial processing is required to correctly form all of the required IMAS relationships.