Study on the extraction method of fine water bodies in the mountainous areas of southern Anhui

Water is one of the most important natural elements in nature, obtaining the spatial distribution information of water plays an important role in people's production and life, flood disaster prevention and other aspects. In this article, taking Xin'anjiang in southern Anhui Province as the research object, combines two kinds of data sources of Landsat 8 OLI image and DEM to extract the small water body in mountain area, by comparing several commonly used water body information extraction indexes, the problem of mountain shadow and part of water body disconnection occurred in water body index extraction is solved by using DEM to extract river network. The specific solutions are as follows: for the study area with light River break, the mathematical morphological expansion filtering and thinning methods are used for processing; For some areas with serious disconnection, the river network extracted by DEM is used to supplement. The results show that the accuracy of water extraction from two data sources can reach 90%, our study has a certain reference value for the extraction of small water in mountain areas.