Digital demodulation algorithm based on frequency correction for the near-infrared spectrometer.

For the scanning near-infrared spectrometer, the employments of traditional digital demodulation algorithms require that the sampling frequency is an integral multiple of the measured signal frequency. However, the measured signal frequency is unstable, and the sampling frequency is a fixed value in the spectrometer, which affects the accuracy of amplitude demodulation results. To improve the signal-to-noise ratio of spectra, this paper proposes an improved digital demodulation algorithm based on frequency correction. Through theoretical analysis, simulations, and experiments, the improved digital demodulation algorithm based on frequency correction was compared with a digital demodulation algorithm based on the average of the absolute value in a complete period. The results showed that the former algorithm decreased the noise caused by the fluctuation of the measured signal frequency. The signal-to-noise ratio of spectra increased from 626 to 1124, and the measurement time increased only 0.73%.