Estimation of relative water content in rice panicle based on hyperspectral vegetation indexes under water saving irrigation

Abstract

Determining grain water content is a traditional method to quantify maturity, but this method is laborious. Ground-based remote sensing with rapidness and flexibility has been widely used in evaluating rice growth, but less attention has been paid to predicting physiological maturity. A field experiment with rice was conducted to investigate the effects of water saving irrigation on physiological maturity based on hyperspectral data. The results indicated that, with higher coefficients of determination (R2 = 0.93 and 0.87, respectively), higher residual prediction deviation (RPD = 3.54 and 2.58, respectively) and lower root mean square error (RMSE = 2.88 and 4.43, respectively) among the tested models, R1654/R662 and R546/R562 were suggested as the optimal indexes for monitoring relative water content in rice panicle under flooding and wetting conditions, respectively. This finding is helpful in providing reference for monitoring rice physiological maturity.