A general theory of spatial relations to support a graphical tool for visual information extraction

In this paper we present a general spatial composition framework which allows one to model the graphical objects and the spatial relations of a large class of visual languages. The new formalism has been implemented within the SRQ tool, a software system for the Visual Information Extraction, enabling it to work on a wider range of domains. In particular, in the paper we describe the application of SRQ to geospatial data.

► We illustrate the theoretical formalization of the contour syntactical model. ► We extend spatial relations to work directly on the graphical objects' images. ► We present the redesign of the SRQ tool to support the contour spatial relations. ► We describe an application of the VIE framework to the analysis of geospatial data.