Study Of Contour Generalization Based On Douglas-Peucker Algorithm

With the research and application of geographic information system continued todeepen, to achieve the map automatic generalization has become the functionalrequirements of the most pressing. As an important component of the map element,contour is the main form of describing landform and contour generalization hasproved to be an important research naturally of map antomatic generalization.The surface morphological diversity of the real world determines the complexityof the contour, which contour generalization with a high degree of difficulty ofscientific research. The current approach of contour generalization can be divided intotwo groups such as points compression methods based on a single curve andgroup-generalization of contour. Douglas-Peucker algorithm, Vertical Distancealgorithm, method of deflection angle and so on belongs to the first class; the secondclass includes structured generalized algorithm, based on the landform evlvation withthe generalization method, etc. The existing method can, to some extent, complete thecontour generalization task, but these methods are far from mature, each with a certaindisadvantages and applicability, and cannot be very good in the process of solvingcontours comprehensive all case.Based on the analysis of existing contour generalization method, a single curvegeneralized method in selecting Douglas-Peucker method as the research object, by alot of effective experiment, papers for some prolems are discussed and improved suchas for from1:10000to1:50000contour generalization process of the threshold ofcertain, self intersection judgement since contour generalized, intersection judgementbetween each contour, and the node recovery at the intersect situation, etc, andproposed base on not fully binary tree topology model of contours according to theactual application requirement; after that, selected group-generalization of contour inthe contour line structured generalized method for comparision, aiming at the featurepoint selection, valley point and ridge point judgement, structure line connection andfeature point front opening judgement and so on makes a brief discussion.Finally, with comparing the two methods experimental results in the same areaand analyzing the advantages and disadvantages of the two methods and the existingproblems, provides more perfect ideas for contour generalization based onDouglas-Peucker algorithm.