Towards a general theory of multi-representation data capture in GIS

Abstract

Normally, spatial databases were built by digitizing existing different scale maps. When there was a requirement for capturing new datasets in order to update databases or build new databases, the process usually follows a similar methodology, whereby different spatial databases were created with representations of each object captured separately. This paper formally describes a multiple representations data capture workflow in which all representations of an object are captured at the same time and result in a multiple representation database or a multiple scale database. The workflow is described using a Unified Modeling Language. A prototype was developed to demonstrate the multi-representation data capture concept. Using the prototype, a database with multiple representation of each object was built within an area containing two object classes – a building and a road network. The result shows that a multiple representation database can be built with the designed multiple representation data capture procedure. Additionally, it was proved that building a multiple representation database with multiple representation data capture concept is less expensive than with the procedure of capturing each representation separately.

Keywords: Multi-representation, Data capture, Object class, Workflow