Prediction of crude protein content in rice grain with canopy spectral reflectance

Abstract

Non-destructive and rapid monitoring methods for crude protein content (CPC) in rice grain are of significance in nitrogen diagnosis and grain quality monitoring, and in enhancing nutritional management and use efficiency. In this study, CPC and canopy spectra in rice were measured based on rice field experiment. Key spectral bands were selected by principal component analysis (PCA) method, and the predicted models were built by multiple linear regressions (MLR), artificial neural network (ANN) and partial least squares regression (PLSR). The results showed that there is a significant correlation between CPC content and key spectral bands. The results of prediction for the three models were in order of PLSR > ANN > MLR with correlation values of 0.96, 0.92 and 0.90, respectively, for the validation data. Therefore, it is implied that CPC in rice (grain quality) could be estimated by canopy spectral data.