Odour sampling system with modifiable parameters applied to fruit classification

The aroma emanating from fruits can indicate the maturity level and thus the quality and shelf life of the products. In this work it is presented an odour sampling system system (called also electronic nose), to classify the aroma of Diospyros kaki, whose working parameters can have variable configuration making the system flexible. The system can be appropriately reconfigured for each type of sampling. It is formed by a metal oxide semiconductor (MOS) commercial sensors matrix, which are individually controlled, being able to modify by software several parameters of their operating point. Working this way there is not a limited number of sensors but a very high number of virtual sensors. The system shows the ability to discriminate aroma of two Diospyros kaki cultivars. Moreover the instrument is capable to distinguish the fruits, inside the persimmon cultivar cluster, according to the ripe state after few days of storage.

► We have developed a new odour sampling system.
► Its working parameters can have variable configuration making the system flexible.
► We have used a sensor matrix where each sensor is individually controlled.
► As an application of the system a classification of Diospyros kaki L. is presented.
► We have used PCA and SOM algorithms in order to study the results.