Non-destructive test to detect adulteration of rice using gas sensors coupled with chemometrics methods

Abstract

In order to accurately determine and evaluate the odour of rice, it is necessary to identify the substances that affect that odour and to develop methods to determine their amounts. For more than three decades, researchers have been studying the factors that produce and influence the aroma of rice. An electronic nose can be used to detect the volatile compounds of rice, while an olfactory machine is capable of classifying and detecting the variety, origin, and storage time of rice with a high degree of efficiency. This study aimed to investigate the efficacy of electronic noses and other chemometric methods such as principal component analysis, linear discriminant analysis, and the Artificial Neural Network as a cost-effective, rapid, and non-destructive method for the detection of pure and adulterated rice varieties. Therefore, an electronic nose equipped with nine metal oxide semiconductor sensors with low power consumption was used. The results showed that the amount of variance accounted for by PC1 and PC4 was 98% for the samples used. Also, the classification accuracy of the linear discriminant analysis and Artificial Neural Network methods were 100%, respectively. The Support Vector Machines method (including Nu-SVM and C-SVM) was also used, which, in all its functions except the polynomial function, produced 100% accuracy in terms of training and validation.