Denoising Technique in Gas Sensing Based on Off-Axis Integrated Cavity Output Spectroscopy

To effectively suppress the system and cavity mode noises in gas sensing based on off-axis integrated cavity output spectroscopy and therefore improve the signal-to-noise ratio (SNR) and gas detection sensitivity, this paper proposes an improved empirical mode decomposition (EMD) filtering algorithm on the basis of the traditional EMD method. In the process of hierarchically decomposing noisy signals, the Savitzky-Golay (SG) filtering algorithm and cross-correlation operation are combined, and a reconstructed filtering signal is obtained by using the filtering signals and correlation coefficients. Simulation and experimental results of methane gas samples show that the EMD-SG filtering method can significantly improve the SNR and reduce the lower limit of gas detection. In addition, compared with traditional wavelet-denoising and Kalman filtering, the EMD-SG filtering algorithm has obvious advantages in processing the Gaussian white noise and the non-linear and non-stationary random noise in the system noise and achieves a better filtering effect. After treatment with the EMD-SG filtering algorithm, the SNR of the absorption signal is increased by 1. 9 times, and the lower limit of the detection is reduced from 8. 7 X 10(-6) to 4.6 X 10(-6). The proposed EMD-SG filtering algorithm based on off-axis integrated cavity output spectroscopy has a high SNR and favorable denoising effect and can effectively improve the detection performance of the system. It provides a new method and a basis for developing low-noise gas sensors based on the off-axis integrated cavity to monitor the atmospheric environment.