Prediction of chlorophyll content of winter wheat using leaf-level hyperspectral data

The leaf-level winter wheat hyperspectral response to its chlorophyll content was examined. Firstly, after the 316 scan line images were acquired, the cube image data was constructed and the region of interest (ROI) was selected, then after the average pixel intensity acquired, using correlation analysis combined with stepwise discrimination method for the origin reflective spectrum and the first derivative spectrum, the optimal wavelengths were selected respectively; the chlorophyll content model using multivariate linear regression (MLR) was constructed based on the above seven optimal wavelengths. After statistical significance testing, three wavelengths were abandoned, and the residual four wavelengths, i.e., 710.85, 767.42, 650 and 520 nm were used to construct chlorophyll content prediction model. The prediction results showed that the determination coefficient were R2=0.8434 and R2=0.7093 for the training dataset and the validation dataset respectively. All of these indicated that with the hyperspectral technology, chlorophyll content of winter wheat could be predicted precisely.