Automatic construction of structured dentritic river networks

Abstract

The automated map generalization is always a challenging work in Geographical Information System (GIS). Regarded as terrain skeleton, drainage system should be considered in research on automated map generalization in the first place. Further, as the most important component of drainage system, generalization of rivers properly becomes a focal point. On land, rivers always connect together in networks, and most of these networks form tree structures. As a set of line features, river networks are generalized from large scale to small scale by two main steps: selective omission and selected rivers simplification. Before these steps, the structured dendritic river networks should be constructed. This paper puts forward an improved method of average direction based on angle of river network connections to detect the river flow and the main stream, and then create a data structure to store the topology relations of tributaries in the river networks. Eventually, the paper gives a comparative analysis with existed methods on river flow detecting and topology relation building.