Hyperspectral assessment of nitrogen nutrition for winter wheat canopy using continuum-removed method

摘要

The hyperspectral reflectance of canopy of winter wheat and data of leaf nitrogen accumulation (LNA) were acquired in primary growth stages under different nitrogen levels in order to monitor winter wheat status and diagnose nitrogen using remote sensing method. A new method was developed to estimate the nitrogen nutrition of winter wheat using continuum-removed method, which generally used in spectra analysis on rock and mineral. The continuum-removed method was effectively used to magnify the object spectral absorption features, and it could be convenient to extract the spectral absorption features. Based on the continuum-removed treatment and the correlation between absorption feature parameters and LNA, results show that LNA increased with increasing the nitrogen fertilization. LNA increased from the erecting stage to the booting stage and decreases from the booting to the heading stage under all nitrogen levels. It is the VNIR regions that were sensitive to LNA. By continuum removal operation, it can be found that the method magnify the subtle difference in spectral absorption characteristics arise from the nitrogen stress on winter wheat. At all stages, total area of absorption peak, left area of absorption peak, right area of absorption peak increased with increasing the nitrogen fertilization, whereas the normalized maximal absorption depth by area decreased. The correlation analysis indicated that all the absorption characteristics parameters of continuum-removed spectra highly correlated with LNA, and the correlation relationship of the whole growth cycle was stronger than that of any single growth stage. But the booting stage is the best at the several single growth stages and the NMAD is the best absorption parameter to monitoring the nitrogen of winter wheat canopy. The range 550 nm to 760 nm are the feature bands for extracting nitrogen information of canopy. The regression analysis on the whole growth period showed that the all regression models between the absorption characteristics parameters and LNA were all extremely significant (P<0.001).Therefore, continuum-removed method is a feasible method for quantificational evaluation of winter wheat LNA.