Extracting Thematic Information Based On Knowledge Discovering

During the establishment of Geography Information System and map database, how to acquire geography information efficiently and accurately is a key bottleneck in the development of GIS. Up to now, there are many methods that can be used in extracting thematic information such as the methods of pursuing line, Mathematic Morphology, Pattern Recognition, Artificial Intelligence, but we still have a lot of unresolved problems in this field such as automatically acquiring of contour spatial information, automatically extracting of complicated symbol from thematic map, automatically distinguishing objects in RS data.To extract thematic information efficaciously, the author tries to take advantages of knowledge discovering to solve some problems. The methods of extracting thematic information based on knowledge belong to Artificial Intelligence and the basic step is: firstly, the visual computer knowledge, mapping knowledge and geography knowledge should be found from the map, then discover according methods to build extracting model.In extracting thematic information based on knowledge, Methods knowledge is the most important affecting the ultimate results. In this paper, the author has proposed some personal algorithms and improved methods directed against special problems including extreme value filter, alternate point pattern recognition, a method of detecting straight line, a quick thinning algorithm, a method of discriminating line width based onangle detecting. These algorithms and methods have successfully applied to concrete precedents such as extracting points boundary regions on forestry thematic maps, classifying contour line on topographic maps.It's necessary for extracting map information to give a friendly and visual interface and a method reservoir. So, during the forming of the paper, the author has made a Visual C++ extracting system named TMINFO, which can afford the following functions: management of project and files, display and inquiry, editing grid data, editing vector data .getting help.Due to the limit of theories and practices, the methods and models in this paper still have some problems .we still have a long way to go before the realization of automatically extracting thematic information without manual interference. On the other hand, through the practices, the author comes to a conclusion that the knowledge discovering is an important orientation in the field of GIS data acquirement.