TDLAS noise reduction algorithm for the raw spectral data under strong interference conditions

Abstract

Tunable diode laser absorption spectroscopy (TDLAS) is a high-resolution infrared laser absorption spectroscopy technique with a non-contact measurement, high spatial and temporal resolution, extensive measurement information, which has been a hot research area at present. Compared to traditional techniques, TDLAS technology has many advantages, but in engineering applications under complex environmental conditions, TDLAS technology is still facing many difficulties. Because of the impact of environmental factors, the measured spectral signal would be distorted, and cannot be used to extract useful information. Therefore, to extract useful information from the raw signal, it is essential to improve the signal to noise ratio. To eliminate interference information contained in the spectral signal, the absorption spectra of the laboratory intends to take data preprocessing methods. In the preprocess, the Empirical Mode Desperation (EMD) method is developed in recent years, which is a new self-adaptive local frequency analysis method. Compared to the method of wavelet denoising, EMD method with adaptive filters is able to achieve a multi-scale decomposition of the noise signal. In this paper, EMD method is taken to eliminate noise and interference signal source decomposition. By reconstructing the actual signal and eliminating the noise components, a better SNR can be achieved.