A Methodology for Spatial Consistency Improvement of Geographic Databases
In any information system the reliability of any results of queries, analysis or reasoning, depends on data quality (positional accuracy, consistency and so on). In some cases, answers cannot be obtained due to a lack of information, whereas in other cases answers are wrong or not complete because of inconsistent data. In geographical information systems (GIS), data quality management has to handle the spatial features of objects, which brings specific problems. The goal of this paper is to describe a methodology for spatial consistency improvement of geographical data sets in vector format. It is based on errors survey and classification. Three kinds of errors are identified which lead to three kinds of consistency, namely structural consistency, geometric consistency and topo-semantic consistency. Each of them needs specific checking and correcting processes. All these processes are integrated in a general framework that is presented in this paper. An application of this framework to the Lyon Urban Community GIS (the SUR) is currently conducted; first results are presented.
Unable to display preview. Download preview PDF.