Estimation of cotton canopy fraction of photosynthetically active radiation (FPAR) and leaf area index (LAI) based on hyperspectral remote sensing data.

Utilizing a quantum sensor and non-imaging hyperspectral spectrometer,2 cotton cultivars 4 level densities planting,in north Xinjiang,multi-temporal canopy hyperspectral data and photosynthetically active radiation data at cotton key growing stages were reoorded,and multivariate regression analysis method was used to analyze the correlated relationship between reflectance and cotton canopy FPAR,LAI,respectively.The results showed that the maximum correlation coefficients between hyperspectral data and FPAR,LAI occurred at 813 nm,758 nm wavelength,respectively;based on the linear regression equation between reflectance spectrum ρ813 and ρ758 and FPAR,LAI,respectively,their correlation coefficients were significant between measured FPAR and estimated FPAR,measured LAI and estimated LAI(rFPAR =0.7199**,rLAI =0.6430**,α=1%,n=70).The regression function accuracies were 96.5%,81.7%,respectively.The maximum value of correlation coefficient between the first derivative spectral data and LAI occurred at 734 nm wavelength,but the correlation coefficient was not significant between the first derivative spectral data and FPAR from 350 nm to 2500 nm wavelength.According to the first derivative spectral value ρ′734 estimate for LAI,measured LAI and estimated LAI was significantly relevant(rLAI=0.6947**,α=1%,n=70),the regression function accuracy was 82.4% and was nearly to the prediction precision of the LAI regression modeling at 758 nm wavelength reflectance.The study showed that canopy hyperspectral data can be used for real-time,nondestructive and quantitative estimate of FPAR,LAI.