Research On Adaptive Map Visualization Aiming At Spatial Distribution Properties

As an important geospatial cognitive means and the visualization result of geospatial research, map changes a lot from design process to technique approach. Instead of simple visualization of spatial data, the research of map visualization focused more on mining spatial distribution characteristics, revealing the pattern of social behavior and explaining the geographical changes. The adaptive map visualization has become a hot topic for the variety sources and forms of spatial data and individuation requirement of map user.Adaptive map visualization is defined that the electronic map or visualization system could alter the properties of itself to adapt the changes of spatial data or environment. The research of adaptive map visualization is carried out in two aspects: map content and map form. Actually, the adaptive map content is the focal point of the research. On one hand, spatial data should be organized hierarchically based on the semantic and spatial distribution characteristic. On the other hand, hidden spatial knowledge could be revealed by the visual language.The main contents of this paper are as follows:(1) The background and significance of the research is discussed, and the advantages of adaptive map visualization are analyzed from the organization of spatial data, the form of visualization and the procedure of cartography. Also, the progress of domestic and overseas researchesis summarized from theoretical study and application, and the shortage that heavy form light content is pointed out.(2) The question 'Adapt to what'is answered from the input and output of cartography. On the one hand, the input of cartography includes the theme, form and other characteristics of spatial data, and this paper focuses on the distribution characteristics of spatial data. In consideration of the scale dependent of spatial data, the spatial distribution characteristics of point feature,linear feature and area feature is analyzed based on macro distribution pattern, meso distributed architecture and micro individual characters. On the other hand, the output of cartography including user information and external environment needs to be considered too.(3)The paper discusses the adaptive organization of spatial data. The granularity of changes in spatial data could be grouped into three levels:element classes, objects and geometrical features, and this paper focused on the latter two based on the LOD model and change accumulation model.First, the LOD model of point feature, linear feature and area feature on the elements level is constructed.? The method of constructing the LOD model of point feature is proposed. By analyzing the spatial distribution characteristics of point group and recognizing the important significance point, the importance of points could be computed. Then based on the Voronoi model, point group should be selected progressively and hierarchical organized. Finally, the scale to display of every point could be set according to square root model.? Take the road network and the river network as examples to discuss the method of constructing the LOD model of linear feature. For road network, we build the Stoke model based on the geometrical and semantic feature, and then hierarchically organize the Stroke objects considering the neighboring distribution relationship and maintaining the distribution characteristics of road network, finally set the display scale of Stroke object according to square root model. For river network, the Delaunay triangulation is constructed to identify the watershed boundary, then, the hierarchical partitioning model can be constructed based on the area of watershed, finally, the display scale of river can be set.? The method of constructing the LOD model of area feature is similar to point feature. In addition, the topological relation between polygons should be maintained.Secondly, the change accumulation model for polyline and polygon is constructed.? The method of constructing the change accumulation model of polyline is proposed. Firstly, the BLG tree of polyline is constructed and the offset of node is adjusted based on the hierarchical feature. Then, the nodes are descending ordered according to the adjusted offset and a linear BLG tree is constructed. Finally, the representation of geometrical features of polyline is accumulated based on the threshold of offset at different scale.? The method of constructing the change accumulation model of polygon is proposed. Firstly, the polygon can be fit with minimum bounding rectangle or convex hull and be subdivided into a serial of'pocket'polygons. The addition or remove of the' pocket'polygons will construct a hierarchical H-Tree. The'pocket'should be arranged by descending order of area and the representation of polygon could be accumulated based on the threshold of offset at different scale.(4)We present some visualization methods to display the macro distribution pattern, meso distributed architecture and micro individual characters, including heat map and kernel density visualization, clustered map and point map. Besides, we discuss the adaptive visualization at different scale.(5) The self-adaptive strategy of map form is proposed mainly from two aspects: map symbol and color. For map symbol, the adaptive strategies of semantic level, geometrical feature and symbol accuracy are all discussed. For map color, the adaptive strategies for various spatial data, different users and complex environment are discussed.