Monitoring Leaf Area Index In Rice And Wheat With Hyperspectral Remote Sensing

Leaf area index(LAI)is an important variable to characterize vegetation canopy structure.The hyperspectral remote sensing technique was recommended as an effective method for non-destructive monitoring.The objectives of this study were to explore optimum vegetation indices and wavelet coefficients for monitoring leaf area index in rice and wheat,and establishing the monitoring models with strong explanation and high accuracy based on a large data set which was compiled by combining in situ measurements from twelve field campaigns representing variations in nitrogen rates,water regimes and cultivars over the whole growing seasons of rice and wheat crops.The results would help to provide technical support for designing spectral sensors in estimation of leaf area index.Firstly,by taking account of the trend basis leaf area index and canopy spectral reflectance in rice and wheat under different plant growth status and planting density,and integrating methods of spectral analysis,crop physiological principles and statistics analysis,soil adjusted vegetation index and differential vegetation index were constructed in spectral coverage from 350 to 2500nm.In addition,the quantitative monitoring models for leaf area index were established with strong explanation and high accuracy in rice and wheat.The results show:the optimum spectral vegetation indices for leaf area index was SAVI(R736,R900)for the best monitor vegetation in the early heading stage,R2 of 0.794and 0.61,SE of 1.16 and 1.17,RMSE of 1.43 and 1.14 in rice and wheat.In the different planting densities,DVI(R741,R866)was the optimum spectral vegetation indices,R2 of 0.733 and 0.857,SE of 1.02 and 0.339,RMSE of 1.19 and 0.33 in rice and wheat in low density.In the different leaf area index value,DVI(R733,R930)was the optimum spectral vegetation indices,R2 of 0.654 and 0.619,SE of 1.355 and 1.1,RMSE of 1.271 and 1.17.Further,a quantitative model of wavelet coefficients and leaf area index under different wavelet functions was constructed to evaluate the feasibility of wavelet analysis on monitoring leaf area index in rice and wheat.The results show that:the accuracy of the model stablished with wavelet features and vegetation index were almost the same.This indicates that wavelet analysis is feasible in terms of monitoring the leaf area index.The best robustness was obtained wavelet coefficient with under gaus2 function,features 7 scale wavelet coefficients at 720nm.In the early heading stage,R2 of 0.729 and 0.662,SE of 1.27 and 1.16,RMSE of 1.359 and 1.132 in rice and wheat.In low density,R2 of 0.725 and 0.861,SE of 1.034 and 0.333,RMSE of 1.238 and 0.329 in rice and wheat.