Detection of Smell Change of Flue-Cured Tobacco Based on an Electronic Nose

In order to make the flue-cured tobacco represent better aroma quality, an electronic nose with an array of metal-oxide gas (MOS) sensors are used to monitor the smell change during tobacco curing process. The design of dedicated electronic nose and experiment steps are described in detail. After signal preprocessing, the smell synthesized curve, of which four obvious peaks were detected, was obtained by the use of principal component analysis (PCA). The results indicate that, the peak time would be of important significance to the smell synthesized curve, which can clearly reflect the overall trend of smell during tobacco curing process. Therefore, back-propagation neural network (BPNN) is proposed to forecast peak time. This paper presents a study on smell change based on electronic nose during flue-cured tobacco process and detection of obvious peaks on the basis of sensor outputs. The law of smell changes obtained from analysis is in agreement with flue-cured tobacco theory.