Hyperspectral characteristic analysis for leaf nitrogen content in different growth stages of winter wheat.

The spectral characteristics in the range of visible light and near-infrared shortwave (400-1000 nm) are analyzed using the ground measured hyperspectral data and leaf nitrogen content (LNC) data of different growth stages of winter wheat, which were acquired in 2013 and 2015. First, the quantitative models for monitoring the LNC at different growth stages of winter wheat were established using the main vegetation nitrogen spectral indices. By analyzing the simulation coefficient of the models, it is demonstrated that vegetation nitrogen spectral indices, which are calculated using these data in this study, should not be an effective quantitative estimate for winter wheat LNC. Second, a method for finding representation wavebands of hyperspectral data sensitive to the LNC of winter wheat is proposed based on the spectral correlation. Using the hyperspectral data, which were acquired in 2015 and the proposed method, the representation wavebands sensitive to the LNC of winter wheat are found. The finding results show that the representative wavebands are mainly located in the purple, red and near-infrared wavelength range, but the representative wavebands are different in different stages. The red edge effects of representative wavebands are obvious. Finally, based on the acquired representation wavebands corresponding to different growth stages of winter wheat, the quantitative models for monitoring the LNC at different growth stages of winter wheat were established using data acquired in 2015 and 2013. The modeling results show that the combination of representation wavebands found to correspond to different growth stages of winter wheat are effective and credible for monitoring the LNC. So, these research results lay the foundation for accurate quantitative monitoring of winter wheat LNC.