基于Delaunay三角网模型在矢量等高线基础上通过弯曲特征分析，给出了一种地形结构特征提取及谷地树组织的方法，与传统的基于DEM或TIN数据结构提取地形特征的方法相比，该方法不仅得到了谷地系统在平面图上的拓扑结构，还获得了河谷在空间分布上的汇水范围，依托矢量线、多边形的几何关系判断，增强了谷地树组织中的结构化，避免了DEM方法中的噪音干扰，方法的核心是Delaunay三角网在单根等高线弯曲深度表达上的层次结构识别。基于系统地实验详细讨论了3种树的结构化组织：等高线弯曲特征嵌套结构表达的二叉树、谷地系统拓扑关系表达的平面结构树、水文意义上主支流河谷表达的语义层次树。

Terrain landform features play major roles in such fields as geomorphology type recognition, relief map generalization, DEM construction and hydrology analysis. This paper presents an automatic method to extract terrain landform features and organize drainage system into tree structure based on bend assessment using Delaunay triangulation model. Compared with traditional DEM or TIN based methods, this pure vector approach obtains not only the topological structure of drainage system in planar graph, but also the valley distribution polygon range. Depending on geometrical computation and judgment of vector line, polygon, the structured properties in drainage representation is enhanced, avoiding the case of noise disturbance in DEM based method. The core algorithm makes use of the ability of Delaunay triangulation in detecting hierarchical structure of each contour line. Three kinds of tree structure organization are discussed: the hierarchical binary tree representing bend inclusion relationship contained in single contour line, the plane structure tree representing valley topological relationship, the semantic hierarchical tree representing valley join level from the point of view of hydrology. This paper gives systemically experiment and detailed comparative analysis.