Previous [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

Journal of Inforamtion Science and Engineering, Vol.12 No.3, pp.345-364 (September 1996)
Modeling GIS Data Using OMEGA*

Chiang Lee and Shyh-Chin Wang
Institute of Information Engineering
National Cheng-Kung University
Tainan, Taiwan 701, R.O.C.

Object-oriented (O-O) data models are known for their rich semantics and modeling power for representing complex data. It is also generally agreed that O-O models can provide more functionalities and semantic constructs for modeling data in applications of geographic information systems (GIS). However, there is a lack of a complete understanding of how to apply O-O concepts in modeling GIS data. In this paper, we propose an object-oriented data model called OMEGA which is designed especially for GIS applications. We discuss in the model the use of O-O concepts in characterizing GIS applications. This model distinguishes geographic data into three major types: geometric objects, geographic objects, and relationship objects. Each of these types of objects is modeled by distinct and proprietary type hierarchies. In the core OMEGA model, there are five semantic association types: generalization, aggregation, (spatial) relationship, geographic constituent set (GCS), and selective aggregation (SA). The major enhancements that have been added to the model include (1) separating of relationships between objects from the object class and grouping of relationships of the same type into a relationship class; these classes can have aggregation as well as generalization associations; (2) a proposed unique set construct, called a geographic constituent set (GCS), especially for modeling of GIS data; and, (3) a proposed selective aggregation (SA) construct which is used to resolve overrefinement problems in a generalization hierarchy. Properties associated with these new constructs are also presented.

Keywords: geographic information systems (GIS), object-oriented data modeling, generalization, aggregation, object-oriented concepts

Received October 11, 1994; revised February 2, 1996.
Communicated by Wei-Pang Yang.
*This paper is supported by the National Science Council under Grant No. NSC80-0408-E-006-13.