Research And Implement Of Automatic Block Generalization Based On Spatial Topology Constraints

Geographic Information System (GIS) has the integration functions of geospatial data acquisition, storage, management, analysis and output, which provides a powerful driving force of the rapid production of digital maps for the cartographer. At present, GIS users have built a variety of geospatial databases. These databases store various thematic information of particular geographic areas, which is as the primary source of information for a variety of applications. With the GIS scale gradually expanding and dealing with the problem increasingly complex, people increasingly realizes the importance of cartographic generalization. The reason is that:on the one hand, based on the principle of conservation and restricted by the limitations of the equipment storage capacity, people usually only build one or few kinds of basic large-scale spatial database of detailed and comprehensive content. On the other hand, people try to rely on the basic database, through multi-thematic extraction, spatial analysis, evaluation and prediction research to the specific region, which will derive various series of thematic or smaller scale database or map. The guiding principle of cartographic generalization is that under the conditions of maintaining the basic structure and characteristics of spatial entities, it will express the content in a clear and legible form. It occupies a very important position in the geographic information system.In geographic information systems, geographical entities can be summarized as point, line, polygon these three types of elements, which have spacial geometric features and attribute features. In the vector spacial database, these entities describe the association between property characteristics and geometric features by topological structure. The link of them is the identification code (ID). In the grid spacial database, the geographical entity is expressed as a pixel or pixel array with pixel ranks number to determine the location, with each pixel value to show the attribute of entity. At present, GIS systems have achieved mutual conversion function between the vector data and raster data, which affords possibility for vector data and raster data mixed cartographic generalization in GIS.At present, the generalization map study is mainly for large scale map, with particular emphasis on the simplification of the graphic shape characteristics, such as the curved shape of the line, the basic characteristics of surface boundary as well as the public boundary treatment. However, in the actual geography space, the distribution of geographical entities, geometry and changes are various. A certain type of map elements can not completely describe all geographic entities, and they only can describe partly. Therefore, to describe the same geographical space we need to use the point elements, line elements and polygon elements, which are in complementary relationship. Current cartographic generalization is always for a single map layer or a single map elements, and in the integrated process, it only consider the single characteristics of map features, while ignoring the spatial relationships with other geographic features in the same layer, such as the topological relationship.On the map of the city, the area load of blocks is about70%. and the generalization results of them directly influence the map quality and the effect of drawing, so a reasonable strategy and a suitable algorithm is particularly important. To achieve blocks generalization not only need in accordance with the principle of geospatial data, keeping the overall morphological characteristics of the blocks with neighborhood roads. In addition to maintaining the consistency of spatial topological relations, to avoid the conflict between the surface features elements causing by the narrow space, such as the intersection phenomenon of roads and blocks. After all generalized operations are completed, we need to assess the quality of the automatic generalization results to check whether they meet the complex requirements of the new scale conditions. On the other hand, we need real-time evaluate the output of the automatic synthesis algorithms and the correctness and superiority of algorithms.This paper mainly research about the block automatic generalization, with the constraint conditions of maintaining the consistency of spatial topological relations, starting from the cartographic generalization research background and current research situation, studying about the principles of blocks automatic generalization algorithm, road generalization algorithm, the resident polygon generalization algorithms and cartographic generalization quality evaluation. At last, based on the platform of MapGIS K10, we experiments on a1:10000block data and it obtains a satisfactory result.The main research contents include:a) Analyzing and summarizing about the principles of blocks generalization, then it presents the problems block generalization faces;b) Combining the constraints condition of the topological relations in blocks automatic generalization, it researches the improved method based on the constraint Delaunay triangular network.This paper improves the traditional method from three aspects, a. Building triangulation inside the road. b. Deleting triangulation of same Z value to avoid dead end lines, c. Further to process crossroads. Finally experimental results show that the extracted axes match original shape and are more coherent in crossroads.c) On the basis of the road centerline extraction, it studies about the adjacent algorithm of polygon and treatment of polygon attributes in the generalization process.d) It researches about block automatic generalization evaluation system, starting from the geometry classification of the geographic entity, presents the quality evaluation system of point, line and polygon elements generalization, in which the boundary points compression ratio, the rate of change of the number before and after generalization, overall morphological similarity and the rate of area change these indicators provide evaluation standards for the quality of the follow-up blocks automatic generalization results.e) Based on the above study, on the MapGIS K10platform, we experiment on block data for automatically generalization. Firstly, it carries out structural analysis towards the original blocks using block centerline, and then under the constraints of the extracted centerline.it focus and realize adjoining, merging and simplified operations to the block units., and finally, according to the quality assessment parameters of line and polygon elements, it gets the quality evaluation of blocks generalization.