Estimating Canopy Biophysical And Biochemical Parametersof Winter Wheat Heading Stage Basedon Hyperspectral Remote Sensing

The comprehensive use of modern agricultural information technology,is directly for the agricultural production process of agricultural technology ideas and practice,is to achieve high quality,high efficiency,low consumption,and environmental protection,the fundamental way of agricultural production.As an important part of the precision agricultural technology system,high spectral technology is becoming more and more important.Hyperspectral remote sensing technology can overcome the traditional chemical analysis method to obtain crop agronomy by virtue of its advantages of large spectral information,high spectral resolution and strong continuity of the band,and can carry out non-destructive and rapid detection of the physiological and ecological conditions and biochemical components of crops Parameter defects.As an important grain crop in China,it is very important to realize the accurate agricultural management of wheat by using hyperspectral technology to obtain real-time information of its growth status.In this study,winter wheat was used as the research object,and the canopy reflectance spectrum and leaf area index of winter wheat were obtained by field experiment and sampling test.The chlorophyll content and nitrogen content of leaves were measured by laboratory.The characteristics of the canopy reflectance spectrum of wheat and the correlation between the spectral information and the measured physical and chemical parameters were analyzed.The sensitive bands and characteristic spectral parameters were screened.The hyperspectral vegetation index was constructed and the different modeling methods were used to establish the physical and chemical parameters Estimate the model,and verify the model.The main results of the study are:?1?The spectral reflectance of the canopy in the visible light?360⁷00nm?winter wheat?heading stage?was low,and the spectral curve had two absorption valleys and one reflection peak,namely 490 nm blue light,680 nm red light and 550 nm green Light.In the short-wave near-infrared is a strong reflection,especially in the 690⁷50nm regional reflectivity rose sharply to form the most important feature of vegetation spectrum.?2?When the relative content of chlorophyll content in wheat canopy is different,the corresponding reflectance spectrum will fluctuate up and down.In the visible light region,the chlorophyll content is higher and the spectral curve is shifted downwards.The reflectance of the infrared high position increases with the increase of chlorophyll while rising.The correlation between canopy spectral reflectance and chlorophyll content in winter wheat was significantly correlated with 400⁷20nm in the visible region,reaching the maximum at 594 nm.The derivative spectra obtained by the first order differential transformation can improve the correlation with the chlorophyll content,but the volatility is larger.Most of the spectral characteristic variables based on visible light and infrared,such as green peak,red valley and "three sides",have a good correlation with wheat chlorophyll content.The estimated model y = 1.196x-844.54,which is derived from the vegetation index REP as an independent variable,can be used to estimate the chlorophyll content of wheat.The index REP has a good correlation with the chlorophyll content of wheat leaves,The preferred index of content inversion.?3?The reflectance of wheat canopy spectral reflectance at different leaf area index levels is significantly different in each band,and the reflectance at the visible light reflection peak near520 nm 580 nm decreases with the increase of leaf area index,while the near infrared platform region The effect of the layer structure causes the reflectivity to become larger as the leaf area index becomes larger.There was a negative correlation between the leaf area index and the reflectance of the visible wavelength band?460⁷10nm?,and the near infrared band at760¹ 000 nm had a significant positive correlation.The model is y = 0.215 x + 0.510,and the coefficient of determination R2 is 0.6604,the root mean square error RMSE is 0.861,and the correlation coefficient is 0.861,and the ratio of the vegetation index to the vegetation index is RVI?698,892?vegetation index.The error RE% is 13.68%,which can estimate the leaf area index of winter wheat heading date.?4?The correlation between leaf nitrogen content and single-band reflectance was negatively correlated.In the visible light band?400⁷20nm?,the leaf nitrogen content was negatively correlated with its reflectance,and reached a significant correlation level between?543⁶40nm?The The correlation between the nitrogen content and the spectral reflectance of the original reflection spectrum is obviously improved,and the empirical value of the selected primitive and first order differential characteristic bands and the selection of nitrogen content is exponentially Independent variables,multiple stepwise regression.Models with RI_1dB,m SR₇₀₅,mNDVI₇₀₅ as independent variables y = 0.84x₁-0.42x₂ + 0.571x₃-0.392 and models constructed with FD₅₀₉,FD₆₈₅,FD₅₃₆ as independent variablesy=-135.97x₁+ 101.43x₂-100.8x₃+ 0.263.The best inversion accuracy is the best model to estimate the nitrogen content of canopy leaf in heading stage of winter wheat.?5?Modeling method by using SVM to these three models for predicting the physical and chemical parameters of the optimization,found that the SVM model in univariate and multivariate model has higher accuracy and stability,is the preferred method for wheat physical and chemical parameters estimation modeling,has great potential application.