关键词: 图形渐变技术, 地图连续尺度表达, 同名特征点

Abstract:

A new Morphing technique for two contours is proposed based on characteristic points matching. This problem occurs frequently during continuous zooming in interactive maps. For corresponding contours at two different key scale maps, firstly, two multi-way trees of characteristic points of these two contours are built, in which the characteristic points on top hierarchy are identified by building convexes of points, which are located on contours and other subordinate characteristic points of tree are detected by using Douglas-Peucker algorithm between two corresponding arcs. So using the method, the similarity of the whole shape of geographic feature at different scales is not only considered, but the influence of extreme points of the curve on curve taken into account as well. Secondly, the different hierarchy characteristic points matching are researched. For each first-level characteristic points of small key scale contour, the most neighboring first-level characteristic points of which on large scale contour are regarded as their candidate corresponding points. Next, the ultimate matching relationship of first-level characteristic points will be established by setting neighboring distance threshold and arcs' shape analysis. For the following-level characteristic points, the matching relationship will be built from higher- to lower-level characteristic points based on distance proximity. Thirdly, the new contour of arbitrary interval scale between the two key scales is acquired using Morphing technique. The mapping line of any point between two continuing characteristic points on large scale contour is built up from the large to small key scale contour according to the distance proportion of the two characteristic points. Then according to the representative scale, some new points can be interpolated on mapping lines and these points can be combined into a new representative scale contour. So in arbitrary scale contour representation can be obtained based on morphing technique. The key point of Morphing technique is recognition of characteristic points and matching relationship of characteristic points. The paper have compared the cost of time in the model provided by this article with that in the algorithm proposed by literature[5]. The conclusion is that the cost of time of the former model is less than that of the latter algorithm. Experiment results show the model based on multi-hierarchical characteristic points is helpful trial of continual-scale representation of contour in the network environment and can effectively sustain topological relationship between contours.

Key words: morphing, map continual-scale representations, corresponding characteristic points