Fourier domain local narrow-band signal extraction algorithm and its application to real-time infrared gas detection

Abstract The local narrow-band signal (LNBS) decomposition algorithm is introduced in non-dispersive infrared (NDIR) gas detection. In order to meet the requirement of real-time detection, the original LNBS algorithm is implemented in Fourier domain to be computationally efficient. The Fourier domain LNBS decomposition algorithm is both time and space saving, so it is suitable for an embedded system. The effectiveness of the technique is evaluated by the measurement of C2H5OH gas concentration. Compared with a digital lock-in amplifier, the Fourier domain LNBS extraction has an advantage of little sensitivity to the frequency drift. Additionally, it has no lagging effect and hence the changes of the gas concentration can be measured in real time.