A correction method for mitigating inter-instrumental absorbance drift in grating-based near-infrared spectrometers

Abstract

In near-infrared spectroscopy analysis, the accuracy of instrument wavelength and breadth is crucial as it forms the foundation for transferring different instrument models. Wavelength and absorbance drift in grating-based scanning instruments result from differences in the grating, detector, and wavelength scanning system. Spectral wavelength drift can be corrected using materials such as polyethylene and polystyrene, which have absorption peaks in the near-infrared region. To reduce absorbance drift in spectra at different wavelength points, this article proposes a point-by-point linear correction method for near-infrared spectra using a small number of typical agricultural product samples. The method constructs a linear relationship model for the absorbance of each wavelength point between the main instrument and slave instruments. The study used seven S450 grating-based diffuse reflection near-infrared spectroscopy instruments, one serving as the main instrument and the remaining six as slave instruments. The point-by-point linear correction method was used to correct wheat spectra collected by the slave instruments, and a crude protein content model for wheat was established for prediction. The results showed that the method reduces spectral differences between different instruments, improves spectral consistency, and reduces prediction errors, achieving better model sharing between instruments. After correcting, the average normalized variation coefficient of wheat spectra decreased by 95.12%, from 7.78% to 0.38%, and the average standard deviation of the predicted results decreased by 78.18%, from 0.5321 to 0.1161. The correction effect of the method combined with traditional pre-processing methods was better than using pre-processing methods alone. Overall, the point-by-point correction method based on a small number of typical agricultural product samples has a significant effect on improving the accuracy of near-infrared spectroscopy analysis.