A Version-Based Modeling Approach for Spatio-temporal Data

Temporal geographic information system (TGIS) aims to collect, manage and analyze dynamic geographic phenomena. In order to represent and process time-evolving geographic phenomena, one of the key problems is to design a good spatio-temporal data model. Much work has been made on spatio-temporal data models or on spatio-temporal modeling approaches, but they have mainly focused on change of spatial data but not of non-spatial data. This paper attempts to address this problem. A novel unified version-based modeling approach for spatio-temporal data is proposed. Based on the conception of version, the evolution of geographic phenomena is represented by a series of object-versions and attribute-versions which are associated with temporal information. This approach processes the spatio-temporal data more effectively and treats spatial data and non-spatial data in a uniform way. Using this modeling approach, the object states can be retrieved conveniently and the evolution history of an object or its attributes can be traced easily.