A Computational Model For Spatial Direction Relations Between Object Groups In Environmentally Geographical Spaces

Environmental protection is one of the major issues of common concern in the humansociety. With the global environment is deteriorating, people are increasingly aware of theimportance of environmental protection and fully recognize the key role played by theinformation technology in protecting environment. Geographic information system forenvironment (EGIS) is a comprehensive technology platform for helping people to protectenvironment. Compared with other information systems, one of the most outstandingadvantages of EGIS is spatial analysis, and the improvement of spatial analysis depends onthe guidance of the theory of spatial relations.Spatial relations are the relations between two spatial objects, which are determined bythe geometrical characteristics of spatial objects. They include spatial topological relations,spatial distance relations and spatial direction relations, etc. Compared with spatialtopological relations and spatial distance relations, spatial direction relations are undoubtedlycloser to people’s lives and have wider practical applications. Their importance has beenproved in the communities of spatial database, spatial semanteme, spatial analysis, spatialquery, spatial reasoning, automatic map generalization and environment geo-spatial cognition,etc.The computational models for spatial direction relations are important theoretical toolsfor describing spatial direction relations between two objects, and are always the ones whichare important and difficulty in the studies of spatial direction relations. At present, the studiesabout the computational models for spatial direction relations between two single objects arerelatively mature. But there are no relevant studies and publications about the computationalmodels for spatial direction relations between two object groups. So the thesis has studiedsystematically the characteristics of the spatial direction relations between two object groups,the main factors impacting the spatial direction relations between two object groups, as wellas the computational model for spatial direction relations between two object groups, takingthe basic principles of Gestalt psychology and relevant algorithms of Delaunay Triangulationas the main theoretical foundation and technical means.The major achievements and innovations of the thesis include the three aspects asfollows:(1) Summarizing the characteristics of the spatial direction relations between two objectgroups into three points: the first is the complexity, that means the process of determiningspatial direction relations between two object groups is the one of complex spatialconsideration and computation, involving many aspects of spatial problems, and its result isimpacted by many factors; the second is the multi-directions, that means the spatial direction relations between two object groups usually can not be expressed by a single direction andonly using multiple directions can be described completely and accurately; the last iswholeness, that means the spatial direction relations between two object groups may be reallyregarded as the relations between two “wholeness” in light of the proximity and closureprinciples of Gestalt psychology, having the characteristic of “the wholeness is greater thanthe sum of its parts”, and can be determined by the spatial direction relations between theirrespective “outline”, having nothing to do with their specific geometric ingredients. It is theimportant theoretical basis to establish the unified computational model of spatial directionrelations for point groups, linear groups, polygonal groups and complex groups.(2) Concluding “the visual area and distribution density are the main factors impactingthe spatial direction relations between two object groups” by analyzing the impacts of somefactors on the spatial direction relations between two object groups, such as the spatial shape,distribution range and distribution density of object groups, as well as the spatial distancerelations and visual area between them. It is the key to establish the computational model forspatial direction relations between two object groups.(3) Based on the above studies, the computational model for spatial direction relationsbetween two object groups is established, and its basic idea is: first, constructing the Delaunaytriangulation of the two object groups, and based on it building the visual area between them;then, getting the quantitative spatial direction relations between two object groups bycalculating the azimuth angles of the rays for indicating directions; finally, classifying theazimuth angles which belonging to the same direction district so as to realize thetransformation from the quantitative spatial direction relations to the qualitative spatialdirection relations.The model proposed in the thesis is realized by means of C#programming language, andthe analyses of the experiments show that the computation results of the model can meetpeople’s habits of spatial cognition and solve the problems about computations of the spatialdirection relations between two object groups in the environment geographic space.