逐层分解选取指标的河系简化方法

Drainage generalization by layered division of the number of retained rivers in river trees

地图综合理论与方法的研究重点之一,是依据上下文环境进行结构化选取和化简。水系作为地图的基本要素之一,在地图综合中必须依据水系的类型和河网密度进行化简,以保持河网密度差异等宏观特征。本文提出了一种基于支流数量差异的选取指标分配与河系简化方法,将选取数量在河流的各个子流域上按比例分配并递归分配到下一层次的子流域,在各个子流域依据长度和间距选取河流,从而实现顾及密度差异的结构化选取。在软件开发基础上进行了选取试验,并对照手工综合的河系图进行了分析。结果表明,基于支流数量差异的比例选取方法切实可行,能够有效地保持树状河系的特点和河网密度的区域差异。

The theoretic and methodological research of map generalization is an important field in cartography and geographical information systems. The difficulties of map generalization lie in selectively omitting some features while maintaining the overall characteristics of their spatial distribution. Simplifying drainages is an important aspect of map generalization. Structural patterns and density differences are salient characteristics of drainages, which should be maintained in generalized versions. This paper proposed a new method to generalize dendritic drainages while maintaining the density differences between sub-drainages, which is based on the layered division of the number of retained rivers in river trees. We analyzed the layered structure of dendritic drainages and their density differences, and found the density differences inside a drainage come from the differences of the number of tributaries of its sub-drainages. A method to allocate the number of retained rivers is accordingly proposed, which divides the overall retained number of sub-drainages of a drainage according to the ratio of the number of the tributaries. The structured drainage generalization is executed in three steps. Firstly, the river entities and river trees are constructed, and the number of tributaries for each river is counted. Then the number of rivers to be retained in the generalized version is computed according to the Square Root Law. Lastly the number of retention is divided among all of the drainages according to the number of their tributaries, the mainstream of a drainage with big length and spacing is selected, and the allocation and selection are recursively executed in the lower layer until the number of the selected rivers reaches the limit of the drainage. The method was implemented as plugins in Java Environment. Experimental results are compared with the hand-made generalized maps. A case study of drainage generalization showed that the method created acceptable results and the density differences among drainages were maintained effectively.