Research Of Generalized Spatial Data Association And Retrieval

The majority of information generated by human is associated with spatial locations. With the generalization of the concept of spatial information, the content of management and retrieval of spatial data is expanding from traditional GIS spatial data types into texts, multimedia, web and other type of data, which contains information of the spatial position. These non- traditional spatial data and those in traditional GIS consists generalized spatial data. With the rapid development of information technology and widely used, the amount of data spatial data continues expanding, how to retrieve spatial information meeting the needs of decisions from these massive, multi-type spatial data has become an important issue of the current spatial data retrieval. The traditional methods based on metadata, spatial area and keyword-words can not meet the current needs of retrieval. Then using the prevalent correlations between spatial information to improve the efficiency of retrieval of spatial data becomes a hot research field.For the key issues in spatial data retrieval techniques based on association, this paper researched association mode of spatial data, creating of association mode and methods of spatial data retrieval, including:Firstly, this paper proposes a correlation model of generalized spatial data for traditional spatial data, place names and general entities. Then we introduce sub-model for association between generalized spatial entities and another sub-model for spatial data based spatial coding. Experiments show that the model supports the efficient retrieval of spatial data well.Secondly, two methods of creating association are proposed: 1) Abstracting feature words based on synthesis support, which implements association between entities. The results show that the effect of this method for the data of organization and characters, places and character types of is good. 2) Creating association between spatial data based on spatial encoding. We get the spatial encodes of spatial data, and then correlate the spatial data and spatial encodes. The result of experiments proved that the method support the data retrieval efficiently.Thirdly, two methods of retrieval are proposed: 1) a method of retrieve generalized spatial data by association based on the spatial area, after spatial encoding of retrieval area, find spatial data nodes and place nodes that satisfy conditions. This method is an intersecting or adjacent retrieval. The results show this method can support related data retrieval. 2) a method of retrieval spatial data based on mode combinations, which supports precisely retrieve.Finally, this paper designed a prototype system for generalized spatial data association and retrieval, which consists of three modules: association creating, associative retrieval and data visualization. By developing a prototype system we verified the model and method proposed in this paper, and provides specific examples for the associated spatial data retrieval application.