Spectroscopic determination of leaf water content using linear regression and an artificial neural network

Abstract

In order to detect crop water status with fast, non-destructive monitoring based on its spectral characteristics, this study measured 33 groups of peach tree leaf reflectance spectra (350 to 1075 nm). Linear regression and backpropagation artificial neural network methods were used to establish peach tree leaf water content and perform quantitative analyses between spectral indices. The results show that a linear relationship existed between the peach tree leaf water content (relative water content and equivalent water thickness) and its leaf reflectance spectral index. The models performed satisfactorily and could be used to detect the water content of the peach tree leaves.