Research On Inversion Method Of Soil Moisture From Vegetated Surface Based On SAR Image

Soil moisture is an important parameter of surface water resources formation, conversion and consumption, which plays an important role in yield estimation and environmental disaster monitoring. So accurately and quickly obtain a large area of bare and vegetation cover soil moisture is an important issue for domestic and foreign research.With the continuous development of SAR technology, the use of multi-frequency, multi-polarization, multi-angle, variable-mode SAR data on a wide range of soil moisture monitoring has made great progress. But now just in bare land area archieve some good results, there is no better way to eliminate the presence of vegetation factors in vegetation-covered area, especially in the tall vegetation coverage area.This paper first analyzes the bare soil and vegetation coverage soil microwave scattering model, and then simulates backscatter change with soil moisture and roughness which is based on AIEM to get simulated data sets, through the analysis of these data, establishes a bare soil moisture inversion model which removes roughness parameters with co-polarization. Using water cloud model to remove the vegetation influence on the total backscatter in the vegetation coverage area to get backscattering coefficient which only contains the bare soil scattering message, then combined with bare soil moisture inversion model, we can establish a soil moisture inversion model in this area.With different application purposes, SAR images’ s processing method is different, and its processing result has a great influence on the precision of the inversion algorithm, so this paper discusses SAR pretreatment processes and methods which are suitable for soil moisture inversion. Before inversing the soil moisture, the text first classifies the features and gets bare soil areas, vegetation coverage areas and other areas before. Then using bare soil areas and vegetation coverage areas inversion algorithms to inverse bare soil areas and vegetation coverage areas, respectively.In this paper the soil moisture inversion algorithm is applied to the pretreated Radarsat-2 SAR data of Dali lake research area in Inner Mongolia to get the distribution of soil moisture. By analyzing inversion values and measured values, we found that the overall inversion accuracy is 73.41% and the inversion precision can meet the requirements, so this algorithm can be used to extract large areas of bare and low vegetated soil water content, and the method of removing roughness effects on scattering coefficient with co-polarization is feasible. This article also uses the inversion algorithm in a terrain fluctuation area in Maoxian country of Sichuan and extracts soil moisture from pretreated Radarsat-2 SAR data of this area, and get that the overall accuracy of inversion result is 69.32%. Although geometric distortion and radiation distortion have been corrected with DEM data, the influence of terrain on backscatter can’t completely be eliminated, and there is also some higher vegetation which has multiple scattering that is not considered in inversion algorithm with soil surface in this study area, so the precision of inversion results than Dali lake area in Inner Mogolia reach is reduced a little with their impact. So the inversion algorithm used in this kind of terrain area, these two aspects can be improved to increase the retrieval accuracy.