| Along with wide and deep application of GIS, the demand for spatial data becomes greater. Because of different operation contents in various fields, the demands on geospatial data are different from each other. According to their particular demand, to get the geospatial data of the same region on the same scale, each department collects them separately, using different data resources, different geospatial data standards, particular data models and particular grading and classing system for spatial objects. This leads not only to great waste of manpower and resource, but also to big differents. So, geospatial data sharing has been a research hotspot of current geographical information sciences. However, the change of demand to geospatial data will definitely bring the change to geospatial data's contents and formats. To meet new demand is the theme of geographical information science development. This paper researched on geospatial vector data fusion theory and method, taking out needful information from different datum, which had different data resource, different data precision and different data model. As can not only save production cost, but also speed the geospatial information updating. It has quite significant meaning to improve current geospatial data quality. Main contents of the paper are as followings:(1) This paper gave out the conception and research category, research actuality and existing problems of geospatial vector data fusion, and built up its framework and flow of data fusion preprocessing. It conclude: data integration, homonymy entity recognizing, geometry data and attribute data fusion.(2) This paper deeply analyzed the fountain of multi-source geospatial data. The objective and subjective reasons were labored, including the self-uncertainties of spatial entity and phenomenon, the limitation of human's cognition, the observing error on spatial entity and phenomenon, the computer's expression limitation on geographical phenomenon. It is quite valuable to geospatial data sharing, integration and fusion.(3) This paper analyzed the main contents of multi-source geographical spatial data integration from three aspects: spatial datum, data model and semantic code. We researched integration modes of four kinds: data exchanging, directly data accessing, spatial data inter-operation and ontology-based spatial data integration.(4) Based on spatial relation theories, this paper put forward geographical spatial vector data matching methods, including sequence matching, bidirectional matching, and parallel matching, solving the spatial entities' matching problems of one-to-one, one-to-many, many-to-one, especially many-to-many. The paper summarized the topological spatial relations math models, also, the geometrical measurement to point, line and area entities; ameliorates the "spider coding" algorithm for point matching; put forward overlap-area-based area entities matching method and algorithm; solved the problems to many-to-many area matching. According to geospatial direction relationships description method, we brought forward calculating method of spatial objects directions similarity with direction-relationships-based matrix models in grid environment, and the calculating models when size and shape similar or not, and the matching results of line and area entities. According to the base theories of ontology matching, the semantic similarity calculating method was raised.(5) This paper take the fusion of digital topographic map and digital chart as an example to validate the spatial vector data fusion model and algorithm raised in this paper.
|
PDF Full Text Request