Bulletin des sciences géographiques de l'INCT
Volume 19, Numéro 1, Pages 2-17
2015-10-31

Optimization Process Of The Extraction And Automatic Vectorization Of A Drainage Network And Its Sub-watersheds : Application To A Watershed Of Oued M'zab, Algeria

Authors : Bessaa Brahim . Belhadj Aissa Mostefa .

Abstract

The characterization and analysis of the com-plex behavior of watersheds has become possible due to the availability of satellite images and/or multi-source aerial images as well as geographic information systems and their development tools. Indeed, the digital form of Watersheds (W) is mainly based on topography, represented by the Digital Terrain Model (DTM). This information is used in the extraction process of a watershed hydrologic network (HN) and the estimation of its geomorphological indices. In this paper, the optimization of the process we have deve-loped is at two levels: in raster mode, we propose a code for intermediate flow directions based on the D8 algo-rithm and a sequential treatment of both single and multi-cell depressions. The automatic extraction of the drainage network is based on multiple thresholds determined by the Upstream Drained Surface depending on the density of the maintained network. The maintained network consists of a selected set of channels highlighting the structural and topographic aspects of the watershed. A Sub-waters-hed (SW) is associated with each of these channels. The analysis and simulation of the watershed behavior rely on descriptive data associated with spatial information, inte-grated into a Geographic Information System; they require the conversion into vector mode. To do this, we propose, at this second stage of the process, a new hybrid algorithm for the automatic vectorization of the hydrologic and SW and the sequential creation of geomorphological indices of the SW. The developed process was tested on the watershed of Oued M'zab (Wilaya of Ghardaia, Algeria) we have used for this purpose, an interferometric digital terrain model of SRTM3 mission.

Keywords

Digital Terrain Model, watersheds, hydrologic network, geomorphological indices, automatic vectorization algorithm.