Quantitative Effects of the Spectral Calibration Accuracy of the Imaging Spectrometer on the Vegetation Red Edge

The red edge position (REP) has been widely used to estimate vegetation parameters and also as a sensitive indicator of vegetation stress. Detailed characterizations of REP can be easily achieved from hyperspectral data obtained from imaging spectrometers. After launch of imaging spectrometers, shift in the center wavelength of the spectral channel may occur due to the vibrations and aging of components. In this paper, we firstly proposed a method to quantify the effects of the spectral shift of imaging spectrometers on the REP. The fundamental basis of the method was the simulation of the vegetation reflectance spectra after spectral shift (SSR). The quantitative relationship between the spectral shift value and the REP derived from the SSR was analysed. The result showed a significant linear relationship between spectral shift value and REP error (R 2 =0.9932). For a 10nm resolution spectrometer, channel center wavelength errors of 10%, 30% and 50% would introduce 0.8nm, 2.5nm and 5.3nm error in REP, respectively. The study indicated that spectral calibration error was an important factor to generate REP shift.