曲线弯曲识别与等高线簇结构化方法

提出基于曲线弯曲识别的等高线簇结构化方法,实现等高线簇中所蕴含的地形结构的提取。首先对基于Delau-nay三角网的曲线弯曲识别方法进行改进,包括伪弯曲的合并、弯曲基点的调整及小弯曲的删除方法;基于此,建立弯曲间的层次嵌套关系及平行相邻关系,实现曲线弯曲的识别方法;然后,根据单条等高线间的空间关系,对弯曲类型进行判断;结合曲线间的空间邻近度计算方法,将等高线簇中的成组弯曲进行提取;最后完成了地形结构的识别,实现了等高线簇的结构化。该方法不仅对地形中的山谷进行提取,并且提取出了其对称结构即山脊。试验结果与水系叠置显示,所提取出的地形结构基本合理,证明了该方法的有效性。

Approach of Curve Bends Recognition and Contour Cluster Structuralization

The paper proposed a method of contour cluster structuralize based on curve bends recognition, and realized the extraction of relief structure which represent by contour cluster. First of all, the method of curve bends recognition based on Delaunay triangulation is improved, which includes the mergence of bogus bends, the adjustment of endpoints, and the deletion of illegal bends. Based on these, a hierarchical and a parallel relationship between bends is constructed, and the method of curve bends recognize is realized. Then, according to the spatial relation between single contours to distinguish the type of bends; based on the spatial proximity calculation method between curves to extract bend sets in contour cluster; and finally realized the structuralization of contour cluster to implement the recognition of relief structure. This method extracts not only the valley but also its symmetrical structure which is ridge represent by the contour cluster. The experiment result overlaid with the drainage system shows that the relief structure extract by the method is reasonable, which proved the validation of the method.