The Relationship Research Between Distribution Of Road Network And Map Initial Scale Based On Rough Set Theory

The interactivity and legibility have been improved greatly from paper map toelectronic map. However, the initial scale design of the electronic map is notreasonable. Map scale remains static and cannot make corresponding adjustmentaccording to changes of surrounding geographic information distribution withpanning. Thus, the users have to do a lot of scaling operation and map readingefficiency is reduced. To solve the problem of setting reasonable initial scale ofelectronic map, the following are studied in this paper.(1) The research ideas and basic theory of rough set. The biggest advantage ofrough set theory is that studying problems starting from the practical instance datadirectly, not requiring expert knowledge, and digging out the implicit and verifiedknowledge. The knowledge is association rules that people can understand, and moreaccord with human cognitive experience. Through evaluating distribution of roadnetwork in the neighborhood space, the relationship between map initial scale anddistribution of road network is studied and corresponding rules between the two isestablished based on advantages of rough set theory. The ultimate aim is to designmap scale driven by distribution of road network and improve map readingefficiency.(2) The method of evaluating distribution of road network. The road network is theimportant element on the map and also is the main part of city architecture. Firstly,node data of road network is extracted and Delaunay triangulations are establishedbased on the nodes. Secondly, the evaluation scope is determined based on spatialneighborhood relation model. Finally, the density of road network is evaluated byspatial distribution index, which is node number in per unit area of nerghborhoodscope.(3) The extracting methods and application of relationship rules betweendistribution of road network and map initial sclale. Firstly, decision table ofrelationship between distribution of road network and map initial scale was ctreated.Secondly, discretization and attribute reduction of dicision table was processed. Finally, extract rules from processed decision table. In order to get rules withgreater coverage, decision rules were excluded whose coverage were less than15%.The distribution space of road network which was not covered by rules was solvedthrough the nearest nerghbor algorithm. The smart initial map scale model establishedon extracted decision rules can output map scale automaticly driven by distribution ofroad network, and reduce users’ scaling operations and improve map readingefficiency significantly.(4) The conclusions were drawn based on experiments. To verify the feasibilityof the method, ArcGIS Engine9.3and Visual Studio2008C#was used to developeand test the experimental system. The experimental map data were taken from Googlemap and the coverage of map was urban and suburban areas of Zhengzhou city. In theexperiment, the main purpose of users was browsing road network, rules of mapinitial scale were extracted based on map scale data in50different locations on themap. The experimental system implanted with extracted rules provided initial scaleand was evaluated by users. The experimental results showed that the smart initialmap scale model adjusted the map scale adaptively and dynamically according todistribution of road network and reduced scaling operations and improved maplegibility significantly.