Research On The Soil Surface Moisture Retrieval Model With Remote Sensing In Horqin Sandy

Desertification is one of the most important factors restricting China's economic development, soil moisture is an important factor in sand ecosystems, which restricts the formation and development of sandy vegetation, but also for monitoring land desertification is an important indicator. With the 3S technology advances, The use of remote sensing tools accesses information on a large area of the surface quickly and easily, allowing detection of a large area of soil moisture. Through the establishment of accurate remote sensing retrieval models can be long-term monitoring of soil moisture sand ecosystem dynamics, for combating desertification has important practical value.In this paper, Landsat TM images for the data source to the western edge of Horqin sandy Wengniute Banner of Inner Mongolia as a research test area. First, remote sensing images with radiometric, geometric correction, atmospheric correction, and pre-treatment, preparing for the quantitative retrieval of soil moisture data. Then uses the thermal inertia information model, temperature and vegetation drought index method (TVDI), BP neural network method to establish regional Wengniute Banner, Inner Mongolia Horqin sandy soil moisture remote sensing information model. Executing Accuracy Test using the theory and practical precision are based on remote sensing information model of thermal inertia, temperature and vegetation drought index method (TVDI), BP neural network method to establish the soil moisture remote sensing information model of the precision test. Statistical results show that: BP neural network method to establish the accuracy of remote sensing information model theory and the actual accuracy of the largest, respectively 81.81% and 92.17%; Temperature Vegetation Dryness Index was established by the accuracy of remote sensing information model theory 76.82%, the actual accuracy of 86.15%; based on thermal inertia method to establish the soil moisture remote sensing information model theory and the actual accuracy of the lowest accuracy, respectively 68.27% and 77.28%. As the surface temperature, soil thermal inertia in addition affected by soil moisture, is also affected by topography, vegetation, soil texture, organic matter content and other factors. From the inversion accuracy considerations, bare soil or sparse vegetation in areas of thermal inertia method should be used to carry out inversion. In the vegetation cover a large area should adopt the temperature of Vegetation Drought Index (TVDI) to inversion of soil moisture.