Predicting grain protein content in winter wheat based on TM images and partial least squares regression

In order to further improve the accuracy of predicting winter wheat grain protein content (GPC) by remote sensing, the study analyzed the quantitative relationship between satellite remote sensing variables and GPC. Depending on the partial least squares regression (PLS), the multivariable remote sensing prediction model and the space level distribution map of winter wheat grain protein content were constructed. For the PLS model, the number of the best principal components was 5, and normalized difference vegetation index (NDVI), structure insensitive pigment index (SIPI), ratio vegetation index (RVI), nitrogenous reflection index (NRI) and plant senescence reflectance index (PSRI) were identified as the sensitive remote sensing variables for predicting GPC. The determination coefficient (R2) and the root mean square error (RMSE) between estimated value and measured value of GPC were 0.642 and 0.307%, respectively. The results indicate that PLS method can provide an effective way to improve the accuracy of predicting wheat grain quality at large scale by remote sensing data.