The Research Of Soil Moisture Inversion Technique Based On Multi Source Data Collaboration

An important basis for the management of agricultural water science is the soil moisture information,such information can provide data support for the estimation of crop water requirement,irrigation water planning,implementation and adjustment of irrigation water,irrigation monitoring,evaluation,improvement of agricultural irrigation water management,greatly improve the efficiency of water use.It is of great significance to improve the efficiency of agricultural water use to obtain information of soil water content accurately.Before the study,the author summed up the current remote sensing inversion of soil moisture content through the literature review.There are modify perpendicular drought index based on red and near-infrared feature space,temperature-vegetation dryness index based on temperature vegetation feature space and passive microwave remote sensing method to inversion soil moisture.In this paper,the method of MPDI and TVDI based on visible light is used to achieve the soil moisture retrieval of five days in the southern part of Hebei province.Based on the inversion requirements,selected HJ-1A/1B,Landsat 8 and MODIS remote sensing data as the soil moisture data source,the spatial and temporal resolution between different data sources are inconsistent,this paper resolved the problem of multi-source data following the cooperative inversion of soil moisture.(1)Under the condition of sufficient high resolution remote sensing image data source,inverting soil moisture in every source data separately and achieve the result of inversion soil moisture.Based on multi-source data feature level fusion theory,fusing the feature of inversion soil moisture from HJ-1A/1B and Landsat-8 remote sensing images,finally achieved have complementary advantages soil moisture inversion results.(2)With the condition of only one kind of high spatial resolution remote sensing image existence,and the data can not cover of the whole study area,then introduction of MODIS remote sensing images of low spatial resolution as a substitute for data of high spatial resolution remote sensing images,and that data is used to cover the incomplete high resolution images.Due to the low spatial resolution of MODIS data,which is not suitable to establish separate inversion model between soil water and drought index,introduced similar time high resolution images can provide NDVI and surface coverage type data for MODIS pixel downscaling basis,it can improve the correlation between drought index and measured soil moisture,then achieved MODIS data and high resolution remote sensing data collaboration of soil moisture inversion results,downscaling the result and mosaic with high resolution remote sensing image based on color balance method,obtained multi-source data collaborative retrieval of soil moisture products.It is proved that the precision can meet the production requirements.(3)At the time of inverting soil moisture,without high resolution data can be used as a reference to support inversion,use of the similar time NDVI and surface coverage data from high resolution image as a reference to improve MODIS data inversion model accuracy.Compared with the uncorrected results,the inversion accuracy is improved obviously.