Potential of spectral ratio indices derived from hyperspectral LiDAR and laser-induced chlorophyll fluorescence spectra on estimating rice leaf nitrogen contents.

Abstract

Accurate estimation of leaf nitrogen contents (LNCs) is essential for nutrition management in monitoring crop growth status. The aim of this study was to compare the potential of hyperspectral LiDAR (HSL) and laser-induced chlorophyll fluorescence (LIF) data in accurately predicting rice LNC. First of all, the intensity values of HSL at 694 and 742 nm and LIF at ~685 and ~740 nm were selected as the characteristic variables to analyze rice LNC using data collected in 2014 and 2015, respectively. Second, spectral indices derived from HSL (only) and LIF (only) were utilized to estimate LNC of rice, respectively. Third, a combined ratio indices (the ratio indices of reflectance to fluorescence and NDVI-based indices at the above four wavelengths) was developed and evaluated in estimating rice LNC. The statistical method of linking these spectral indices to rice LNC was the artificial neural network, which was to obtain the optimum performance in LNC estimation of rice. The results demonstrated that the combined ratio indices, especially the ratio of reflectance to fluorescence at ~740 nm, showed a moderate relationship with rice LNC (R2 = 0.736, 0.704, and 0.713 for the 2014 first experiment, 2014 second experiment, and 2015 experiment, respectively).