Monitoring Chlorophyll Fluorescence Parameters At Canopy And Leaf Scales On Hyperspectral Reflectance In Wheat

Chlorophyll fluorescence parameters are usually used to detect the photosynthetic system function,plant nutritional status and the effect of environmental stress.Chlorophyll fluorescence signals were sensitive to the change of environment,so it can explain the principle of the physiological and biochemical process.As the wheat leaf is the research object,this research extract the spectral features adopting the vegetation indices and continuous wavelet analysis to quantity the relationships between the spectral features and the wheat leaf Fv/Fm based on the field trials which contain the treatments of different cultivars,density and nitrogen levels in two growing seasons and monitor the biochemical and physiology dynamics.The research provided an advanced method and solid theoretical basis for monitoring the crop growth rapidly,accurately and nondestructively.Firstly,on the basis of the coefficients of the wheat leaf chlorophyll fluorescence parameters and different physiological indices,the chlorophyll fluorescence parameter Fv/Fm was selected,which is sensitive to the photosynthetic indices,pigment content.The relationships between the leaf Fv/Fm and vegetation indices were analyzed systematically.With the validation of the independent data sets in 2016 season and LOOCV,the precision of the inversion of leaf Fv/Fm by D705/D722(R2=0.3487,RRMSE=1.39%)is the highest at the leaf scale.At the canopy level,R440/R690,D705/D722,D??/D744 can monitoring the wheat canopy leaf Fv/Fm more accurately than other vegetation indices.All in all,no matter at the leaf scale and canopy scale,the prediction effect of wheat leaf Fv/Fm by D705/D722 was good.Moreover,the time series change of the vegetation indices which is significantly related to the wheat leaf is similar to the trend of leaf Fv/Fm.What's more,the more similar to the trend of the Fv/Fm,the highly related to the leaf Fv/Fm.Besides,due to the shortcoming that the vegetation indices which only use several bands,and can't reflect the information during a spectra region,the continuous wavelet analysis was used to capture the small change of spectra shape within a certain wavelength range in this study to extract the wavelet features of which the relationship with Fv/Fm is on the top 1%.At the leaf scale,the wavelet features were mainly distributed around 702 nm in the fourth dimension.Meanwhile,the wavelet feature in 763 nm and 687 nm band is highly related to the wheat leaf Fv/Fm.With the validation of independent data sets acquired in 2016 season,the precision of CWT(4,704)(R2=0.5100,RRMSE=1.21%)to detect the wheat leaf Fv/Fm is the highest,which was higher than that of all the vegetation indices.At the canopy scale,the top 1%wavelet features were mainly distributed around 726 nm in the seventh dimension.With the validation of LOOCV,CWT(7,726)(R2=0.6467,RRMSE=1.38%)can monitor the wheat canopy leaf Fv/Fm more accurately than all the other wavelet features and vegetation indices.However,763 nm and 687 nm bands were not included in the top 1%of wavelet feature.All in all,continuous wavelet analysis can monitor the wheat leaf Fv/Fm more accurately than vegetation indices both at the leaf and canopy scale.This research build the models to estimate the wheat leaf Fv/Fm at different scale by two data processing methods,which provide new technique to detect the chlorophyll fluorescence signals and the plant physiological and biochemical status,offer reference for the bands selection of instrument development to track chlorophyll fluorescence signal,and provides a solid theoretical support for crop precision management at a large scale at the same time.