Signal analytical processing based on wavelet transform for tunable diode laser absorption spectroscopy

Gas analysis based on tunable diode laser absorption spectroscopy (TDLAS) provides features of high sensitivity, fast response and high selectivity. When target gas concentration is below a few parts-per-million spectrometers become more and more sensitive towards noise, interference, drift effects and background changes associated with low level signals. It is purpose of this paper to address some of the problems which are encountered at this low signal levels and to describe a signal processing strategy for gas monitoring with wavelet transform. Different parameters of wavelet by taking gas detection of H2S as an illustration are studied and an improved wavelet-based signal enhancement process is proposed based on the feature of TDLAS second-harmonic signal. The algorithm uses bior3.9 wavelet basic function and multi-resolution decomposition. The results show a plausible improvement in performance of TDLAS system and enhancement of detection limit from 10ppm to hundreds of ppb level under various noise conditions.