A basin recognition method by landform classification and geometrical feature discrimination

Abstract

Aiming at the geomorphological and geometrical features of the basin, a recognition approach combined with a landform classification and discriminant analysis is presented. First, a k-means algorithm is applied to the digital elevation model data. The basic types of landforms with elevation, slope, and relief amplitude are clustered, and the landform regions of plains, mountains, and hills have been separated. Focusing on the plain and its neighboring region, a measurement function is defined to analyze the cases of the plain surrounded by mountains and hills. Finally, a discrimination model for basin recognition is presented with parameters of the surrounded state, area of the plain, and difference of its elevation. The approach is applied to the DEM of the Shuttle Radar Topography Mission of China. The experiments show that the results are the same as those of real terrains.