Integration of imperfect spatial information

The theme of this paper is integration of information arising from observations of spatial entities and relationships. The assumption is that observations are imperfect; in particular, that they are imprecise and inaccurate. Each observation is made in a context that among other things provides a level of resolution. So, a treatment of integration of observations of this type must take account of multiresolution spatial data models. After an introduction, the paper discusses an ontology of imperfection, focusing on imprecision and inaccuracy. The paper goes on to consider logics that are appropriate for integration of information arising from imperfect observations. Two case studies, showing some of the facets of this treatment are developed in greater detail. The first case study considers integration of imperfect (inaccurate and imprecise) observations of a single spatial region. The second case study develops the theory of regions with broad boundary to address the issue of integrating imprecise observations of spatial relationships.