Analysis Of Wheat Hyperspectral Characteristics And Estimation Of Main Physiological Parameters Under The Waterlogging Stress

As one of my country's three major food crops,wheat production plays a vital role in national food security.In the context of climate change,extreme weather and climate events such as waterlogging disasters are increasing,and the frequent agro-meteorological disasters are increasingly affecting agricultural production,seriously affecting food security.As a fast and non-destructive monitoring technology,hyperspectral remote sensing technology can obtain plant leaf information and has been widely used in crop production.Using remote sensing technology to obtain timely and accurate crop growth physiological parameters to grasp crop growth and moisture status is of great significance for timely waterlogging and waterlogging reduction and crop waterlogging assessment.In this study,a pot experiment was designed to treat winter wheat with different gradients of waterlogging stress at the jointing stage.The spectrum data of winter wheat leaves,leaf water content and chlorophyll content data were collected.The data was collected every 7 days after the start of treatment until the wheat matured;analysis of winter wheat the relationship between growth physiological parameters and spectral characteristic parameters,sensitive spectral indices,different pre-processed spectra and sensitive bands,and linear and non-linear regression are used to construct the inversion estimation model of wheat growth physiological parameters,spectral characteristic parameters and vegetation index,and compare them Accuracy and stability of models constructed by different modeling approaches.The research results of this article mainly include the following aspects1.Different waterlogging treatments had obvious effects on the physiological parameters of wheat growth,and all parameters showed a gradual decreasing trend with the increase of waterlogging days.In addition,the spectral reflectance of winter wheat leaves and the first derivative spectral reflectance under different treatments were significantly different.The spectral reflectance of leaves decreasesed continuously with the increase of leaf water content and leaf SPAD in the visible light range,and the opposite was true in the near-infrared region.The first derivative spectral reflectance had the most obvious difference in the red edge region.2.The vegetation index SAVI,hyperspectral characteristic parameters AR,Rg,Rr,Db,SDb,SDy,SDg,SDr/SDb,SDr/Sdy and SPAD values are highly correlated.The vegetation index NDVI and NDWI and the new vegetation index RVI(505,526)had a good correlation with the leaf water content of waterlogged wheat.3.Applied linear and nonlinear regression modeling methods to construct the inversion estimation model of the spectral characteristic parameters of wheat leaf water content and SPAD value and the vegetation index.Through the calibration test,the spectral characteristic parameter AR was better to monitor the SPAD value under waterlogging stress,and the polynomial model constructed by the new vegetation index RVI(505,526)had a better effect on predicting the leaf water content.The inversion model of winter wheat leaf SPAD value and leaf water content established by BP neural network had the highest accuracy.The estimation models of wheat leaf water content and chlorophyll content were based on the BP neural network model with the highest prediction accuracy,the smallest error,and the most stable model.