Non-differential dual-channel mid-infrared absorption spectroscopy for CH4 detection using Wiener-Homomorphic filtering technique

Abstract

A non-differential dual-channel mid-infrared methane (CH4) sensor system was developed by employing a dual-source dual-detector structure. The reference channel operating at the same wavelength of 3.31 μm as the detection channel was used to derive the noise in the electrical and optical domain. The obtained noise was supplied to a Wiener-Homomorphic filter as a reference for suppressing the additive and multiplicative noise in the polluted signal output from the detection channel. Numerical simulation and experiments were performed to verify the function of the Wiener-Homomorphic filtering algorithm by directly applying noise into the sensor system. Under the intrinsic noise level of the sensor, the measurement precision was improved from ~75 parts-per-million (ppm) to ~50 ppm with a ~3 s averaging time due to the use of the technique; with external applied optical interference at 3.31 μm into the system, the technique can well correct the sensor drift and improve the precision from ~150 ppm to ~100 ppm. The proposed technique has the ability of suppressing additive/multiplicative noise in the electrical and optical domain and can be effective in noisy environment for CH4 detection.