| How to abstract and represent the characteristics of geographic information in a systematic and standard form which can be directly understood by computers has been one of the significant theoretical issues on geographical information science. The challenge is how to combine the basic ways for solving problems by computers and information requirements for solving geography problems. As the basic composing element, the formalization of description patterns of spatial location information could be the best way to compute the geographic information issues. This project proposes a creative idea by the introduction of Formal Language Theory to realize formalized representation of spatial location information. In this paper, there is a main technical idea of "characterization-modelization-formalization". The main research contents and conclusions include the following aspects:(1) Construction of description patterns of spatial locations in natural languageThrough enriching and improving the definition and connotation of spatial locations, basic elements of spatial toe at ion are analyzed and then a classification system of spatial to cations is constructed. Based on this classification system, the scale dependence, spatial-temporal features, ambiguity and variety of spatial location are deep studied. In addition, from the perspective of basic components of spatial location,28 kinds of description patterns of spatial locations in natural language are systematically generalized by a combination of texts and graphs based on the large-scale tagged corpus and description of spatial locations, Which can transform the spatial location expressed in text into one described in formalized and structured way.(2) Formalized expression of spatial locations in formal languageBased on the theories of formal language (e.g., first-order predicate logic and description logic), a formalized definition of spatial location is proposed in this paper. Description logic system is used here to describe the features of spatial objects, spatial relationship, attribute information, emporal information, etc., in spatial locations. Then, predicate logic of first order are taken to formalized the expression of spatial locations comprised of different description patterns, and thus, formalized an multi-expression of spatial locations in ways of "feature modeling", "quantitative and qualitative approaches" and "static and dynamic methods". In addition, taking Lu Mountain as test area, the pattern types of description content of spatial location is divided and each pattern type of spatial location descriptions is transformed into its corresponding formalized description and thus, spatial location can be described in an integration method-text, formalization and graphics.Results above show that there are significant differences in spatial location description between in natural language and in GIS, such as in spatial referencing, geometrical shapes, positioning, ambiguity, etc., which are elements of spatial location. In addition, through the formalized description of spatial location in natural language, it is beneficial to systematically and deep interpret the rules of human recognition and expression. Especially, the large-scale corpus provides complete data for the formalization of spatial location. Furthermore, it’s shown that formalized language not only provides uniform mathematical semiotic systems and rigorous logic expression system for multi-level formalization of spatial location description, but also performs perfectly in describing the variety of spatial locations and their complex relationship, which provides base for deep research on formalization of geographic information. |
PDF Full Text Request