Formalism for representing data quality in non redundant spatial information

The geographical information sources, resulting from a collection of different imperfect observations, often contain data redundancy. In this case we need to reduce the redundancy to obtain a non-redundant source before using it. Moreover, a geographical information source may contain data quality information, what means that the data source associates to any piece of information (geographically referenced), a value in the domain of one or several quality components. Geographical information pieces can sometimes overlap hence there is an issue on how to characterize the quality of the information within the intersection. In this paper we consider the case, frequent with geographical information, where the semantic information domain is structured in a lattice, which is a particular case of hierarchy. Therefore, spatial overlapping can lead to redundancy or to conflicts. We present a formal background to reduce the redundancy and to answer questions about the assessment of quality when concurrent spatial information is combined within a common lattice structure. An algorithm to detect and reduce the redundancy, and to compute the resulting quality, is shown with an application example.