Wireless and portable sensor system to differentiate musts of different grape varieties and degree of grape ripeness

Abstract

There are a big number of studies about wine regarding the vinification process and the final product (measuring defects, aromatic qualities, etc.)· But in these studies the must, grape juice prior to fermenting, are mostly ignored. In this paper we report two novel applications of a portable and wireless sensor system (e-nose) for recognition and classification of musts with respect to different ripening processes in the wine producing industry and also with respect to different grape varieties. These applications are very interesting because the musts smell little and in a very similar mode so it is very difficult to distinguish them by a sensory panel. Therefore the system could be used to monitor the evolution of the different types of musts, and to asses some of their characteristics such as the grape variety. We have used several statistical methods to analyze the data such as Principal Component Analysis (PCA), Probabilistic Neural Networks (PNN) and Canonical Correlation Analysis (CCA).