Research on regional soil moisture dynamics based on hyperspectral remote sensing technology

Abstract In order to realize the dynamic analysis of regional soil moisture content in the upper reaches of Xiashan Reservoir in Weihe River Basin, Weifang City, Shandong Province, a regional soil moisture dynamic analysis model based on hyperspectral remote sensing technology is proposed. The process of predicting and deducing the soil water content through the spectral information of remote sensing images, surface parameters and other data can reflect the soil spectral information. The band sensitive to the change of soil moisture in remote sensing image is directly extracted, which is used as the independent variable input in the soil water logging inversion model, and the process of soil moisture prediction is conducted with the measured soil moisture value as the dependent variable input. The SI-1, SI-2 and other water spectral indexes that can reflect the water information are constructed, and the water spectral index used for soil water content retrieval is constructed through various relationships between bands. Compared with the spectral reflectance alone, the use of water spectral index greatly improves the retrieval accuracy. The particle filter algorithm is used to assimilate the observed data and simulated data to obtain the assimilated soil moisture. Error evaluation indicators are used to evaluate the accuracy of the obtained assimilation results and further use different particle numbers and observation errors to conduct multiple data assimilation experiments to explore the sensitivity of particle numbers and observation errors to the assimilation results. The root mean square error (RMSE), relative error (RE) and mean absolute error (MAE) are used to test and verify the effect of data assimilation. The test shows that the dynamic analysis of the regional soil moisture content in the upper reaches of Xiashan Reservoir in the Fangwei River basin by using this method the distribution map obtained by dividing the vegetation coverage are more consistent with the actual situation in terms of the proportion of different soil water salinization degrees, the regional distribution of water, etc. The idea of using hyperspectral remote sensing to retrieve soil moisture is effective.