Research On Spectral Characteristic Analysis And Prediction Model Of Soil Moisture In Songnen Plain

Land plays an irreplaceable role in the history of human existence and development.Under the background of the scarcity of land resources,precision agriculture technology is the inevitable trend of agricultural development.Precision agriculture requires dynamic,real-time,macro and quantitative monitoring of the related information of land,and hyperspectral remote sensing technology can meet this requirement.Soil moisture is an important standard to measure the physicochemical properties of the soil,it is particularly important in the field of agriculture,soil moisture is not only the important factors for crop growth,but also an important parameter in the management of irrigation,and therefore,soil moisture monitoring has been a concern all the time.Traditional soil moisture monitoring methods need lots of time and strength,it's a waste of resources.And the hyperspectral remote sensing technology based on soil reflectance spectral characteristic has advantages of the real-time dynamic monitoring for the soil moisture,which can provide information support for precision agriculture and protection of land resources.In this paper,took Songnen Plain in Heilongjiang Province as the study area,researches on the black soil,chernozem,meadow soil and aeolian sand soil.The research was mainly divided into two parts,the first part studied with a single type of soil-black soil as the research object,measured spectral reflectance in lab,extracted reflectance spectral characteristics parameters by continuum-removal method,analyzed the relationship between soil moisture and spectral characteristic parameters,tried to establish a hyperspectral prediction model,in order to predict the potential of soil moisture by the application of spectral characteristic parameters,and provided a new method for rapid determination of soil moisture.The second part would take a variety of soil types in Songnen Plain as the research objects,data processing methods divided into continuum-removal method and differential processing method these two,tried to establish a unified soil moisture prediction model for different soil types,so as to provide a reference for monitoring soil moisture in the research area.The conclusions are as follows:(1)The absorption valley near the 1920 nm band could be used as the characteristic absorption valley of soil moisture for black soil;predicted the black soil moisture by the spectral characteristic parameters.The unary linear regression model,which was based on the absorption valley area near 1920 nm band,was of high accuracy in prediction result and less input,it could be used as a theoretical basis for the development of soil moisture measurement instrument.(2)Take a variety of soil types(black soil,chernozem,meadow soil,aeolian sandy soil)as the research objects;the soil moisture prediction model differential processed by reflectance hadthe best stability and the strongest prediction ability.While analyzed the relationship between absorption valley spectral characteristic parameters and soil moisture,we found that the absorption valley spectral characteristic of aeolian sandy soil was very different from the other three soil types,in a soil moisture prediction model,with aeolian sandy soil removal,the accuracy of the model would be greatly promoted.In conclusion,we discussed the possibility of using soil spectral reflectance to predict soil moisture through research,and had gotten good effects.It provides the theoretical basis and technical support for the hyperspectral remote sensing measurement of soil moisture.Hyperspectral remote sensing has high spectral resolution,which can quickly obtain the information of soil spectral reflectance,what is widely used in related researches of soil physicochemical parameter prediction.Soil moisture prediction by hyperspectral remote sensing technology provides a new technical means of monitoring the soil moisture in dynamic monitoring area or blocks,and also provides technical support for better protection and utilization of land resources and land use structure adjustment.