Automatic Technology For Map Generalization Based On Web Mapping

The rapid development of networking technology is driving GIS to serve traditional application through the Internet. And thanks to the open geospatial data published by OpenStreetMap, web mapping is now the most popular way for map making. Meanwhile, geospatial data tends to be bigger in size and higher in precision as a consequence of the fast developing technology of geo-relate sensor. Rendering big geospatial data is a time-consuming work and therefore the frequently re-rendering of map by web mapping application will cause a big delay. This problem becomes more serious when vector data is considered. Thus, the visualization of big geospatial data based on network framework becomes a hot topic in GIS researchers.Focusing on vector data, this paper tends to solve the problem by methodology of map generalization. By generalizing the vector data, redundant data is reduced and therefore speeds up the progress of rendering. More specifically, this paper contains the following works.Firstly, a pyramid model for vector data is proposed. By partitioning vector data in different scale into blocks, this model can not only return spatial data on-demand, but also serve as spatial index. More importantly, the model can preprocess the vector data before it is stored which plays a key role in accelerating render speed and offers a good storage structure for the result of generalization.Secondly, proposing a line generalization method based on lines' importance in visualization. If a large number of lines are drawing in a limited map region, redundant line will surely be caused which can be effectively removed by the method proposed to improve the performance of line rendering.Thirdly, proposing a real-time generalization algorithm for point data based on quad-tree structure. By combining the fast operation for point generalization with the quad-tree structure, which is naturally good at organizing multi-scale data, the algorithm is able to solve the spatial conflict among points in real-time.Finally, designing web mapping application to demonstrate the validation of the methods mentioned above.