Skip to Main Content
Considering uneven distribution of watershed land surface in space is an important characteristic feature of distributed hydrological model. The development of remote sensing, in particular the digital technology, provides technical support for solving the problem of uneven distribution of rainfall and watershed land surface in space. The information of vegetation, land use, soil and geology can be extracted from the remote sensing data for watershed land surface. This paper focuses on extracting the data layer of land use from remote sensing image for distributed hydrological model, namely soil and water assessment tool (SWAT). It extracts the data layer of land use from remote sensing image for the SWAT model of the Jinsha River in district of Panzhihua - Huili - Huidong. The methods and procedure is as follows: (1) Preprocessing the raw image data of remote sensing (2) Selecting the appropriate band 741, synthesizing image of remote sensing data, producing the synthetic image of remote sensing for the study area. (3) Making the land use classification map of remote sensing image by imagery processing and classification of remote sensing. (4) Converting land use classification maps of remote sensing into raster data of ESRI Grid format in the ARCGIS platform. (5) Outputting the re-classification maps of land use by applying the code of SWAT model. The result is that the data layer of land use classification for ESRI Grid format was established in the watershed of study area. The types of land use are nine categories. Contrasting the classification map of remote sensing and ground truth region of interest(ROI), the total accuracy of classification and Kappa coefficient are 79.4%, 0.75.