Highlighting of Hydrographic Objects on Satellite Images for the Development of High-quality Digital Relief Models of Floodplain Areas as a Basis for Modeling the Hydrological Regime

Mathematical modeling of surface water dynamics allows us to make forecasts of the hydrological regime of a territory for a wide variety of hydrological, environmental, and geophysical applications. The such simulations quality is largely determined by the relevance of the topographic data used to construct digital terrain models. The important task is the identifying of the boundaries of numerous reservoirs in a floodplain area. The distinctive feature such area is the complex structure of small channels, vegetation, and wetlands. We develop the method for detecting coastlines using multi-channel satellite data which includes two infrared bands (4 and 5 bands), such as Landsat ETM+. Software was tested whith the Northern part of the Volga-Akhtuba Floodplain (Russia) and one of the region in the Amazon basin. It differs significantly in their natural characteristics. During the software testing we determined the conditions that provides accurate detection of borders of water bodies.