Investigation of electronic nose system in classification of five types of native medical plants from Lamiales order

The re-orientaion of human to nature and natural products has led to the development of pharmaceutical, cosmetic and food industries based on natural products. Therefore, the demand for raw materials has been increased, and one of the big batches of these raw materials are medical and aromatic herbs. One of the important issues in large pharmaceutical factories when preparing raw herbs is the differentiation and classification of different plants with similar sensory characteristics. In this study, the classification of five types of medical plants from the Lamiales order was investigated using electronic nose system based on metal oxide semiconductor (MOS) Sensors. After preparing the plant samples, the response of the system sensors to each of the tested plants was recorded and then by principal component analysis (PCA), linear discrimination analysis (LDA) and artificial neural network (ANN) was examined to classify these plants. Resault of principal components analysis (PCA) with using of electronic nose system data revealed that the two first main components cover the 95 percent of data variance. Accuracy of classification using electronic nose for LDA and ANN methods were obtained 92 and 100 percent respectively. The developed electronic nose system has succeeded in classifying the medical plants and could be used as a sensitive, reliable, and fast replacement for traditional methods.