The Validation Of Soil Moisture Remote Sensing Products And Spatial Scale Conversion

With the continuous development of remote sensing technology and applications, there have been more and more remote sensing data with different characteristics of different types o f sensor parameters obtained, how to solve arguments between multi-source remote sensing data normalization, how to convert different scales of remote sensing data to verify the authenticity of the domestic satellite remote sensing products, and some other issues, is the key to push the further development of remote sensing advance.In this thesis, the full text use soil moisture as the main line, through upscaling verification between adjacent spatial scale levels to achieve information circulation and convert in different kinds data, such as ground data, high-resolution remote sensing data products and large-scale low-resolution remote sensing. The main research contents and results include:(1) Through the scheme of the system on the ground and humidity data of the study area and a sample collection. Products based on remote sensing pixel geometry information on the coverage of the measured data to classify statistics,and the conclusion of get optimal number is 8.(2) Ground resolution remote sensing data validation product. HJ-1B use of remote sensing data utilization temperature vegetation dryness index, been characterized TVDI humidity of soil moisture products. Mean measured data corresponding to the pixel value correlation TVDI verification, fitting a negative linear correlation coefficient R2 = 0.6979 has a good linear relationship with the theoretical results.(3) MODIS derived products under the same resolution on the domestic FY-3B inversion verify the authenticity of the product.Performing spatial parameter normalization process, using three resampling approach to image matching between cells and the results were compared. Bilinear interpolation verification result is 12.5661% relative error, than the cubic convolution method to verify the results of 13.754% and 14.0354% of the nearest neighbor method. Bilinear interpolation, cubic convolution, nearest neighbor correlation coefficient R2 are three ways to verify the results were 0.964817,0.957623,0.955537, showed a high degree of correlation.(4) High resolution inverse moisture product validation for medium-resolution remote sensing products.Obtained mean relative error was 18.5646%, the correlation coefficient R2 = 0.948119. And then combined with two different types of surface area of the city and forest(suburb) were compared, the forest area of 16.6357% relative error is less than the average picture of the whole image mean 18.5646%, and because of the urban 37.0241% relative error.(5) The text on the basis of previous research, further from the object-based, pixel-based, the relative error of upscaling conversion obtained results show the impact of factors were explored. In NDVI threshold verification results were summarized and analyzed, and did not reach the level of verification errors NDVI significant correlation. Obtained before sampling the size of the area to verify the relative standard deviation magnitude of the error showed the correlation coefficient R2 of 0.6134.In summary, this thesis in the experiment achieved product verification humidity levels between different scales, to overcome a certain extent on the scale effect problems in the validation process, and through simple and effective indicators of the results were evaluated and analyzed. Thesis research ideas and experimental results for the field of multi-source remote sensing data assimilation, made the validation of satellite products and Spatial Scale Effect of certain theoretical significance.