|
Development of a cost effective customised electronic nose useful for discrimination of tea quality |
|---|---|
| รหัสดีโอไอ | |
| Creator | Hiramoni Khatun |
| Title | Development of a cost effective customised electronic nose useful for discrimination of tea quality |
| Contributor | Debashis Saikia, Utpal Sarma |
| Publisher | Asia-Pacific Journal of Science and Technology |
| Publication Year | 2566 |
| Journal Title | Asia-Pacific Journal of Science and Technology |
| Journal Vol. | 28 |
| Journal No. | 5 |
| Page no. | 9 |
| Keyword | E-Nose, MOS, PCA, Radar plot, Tea taster, VOCs, OLR |
| URL Website | https://www.tci-thaijo.org/index.php/APST |
| Website title | https://so01.tci-thaijo.org/index.php/APST/article/view/255384 |
| ISSN | 2539-6293 |
| Abstract | This paper presents the development of a cost-effective electronic nose (E-Nose) system using five commercially available metal oxide semiconductor (MOS) sensors- TGS2600, TGS2602, TGS2610, TGS2611 and TGS2620. The developed E-Nose system was used to evaluate the quality of different black tea samples. Here, fifteen different tea samples were taken from different tea gardens for testing using the E-Nose and the Overall Liquor Rating (OLR) is obtained from the tea-taster. Further, the data collected and recorded from the E-Nose were analyzed using three machine learning techniques viz. Principal Component Analysis (PCA), Radar plot and K-Nearest Neighbors (KNN). PCA shows the variance discrimination rate of 98.77% and 96.15% for dry and infused tea samples respectively. Radar plots show the sensor TGS2620 as the most significant, whereas TGS2611 is the least significant and the sensors TGS2600, TGS2602 and TGS2610 show almost equal significance. The classification accuracy obtain from the KNN classifier is 97% and 94% for dry and infused samples respectively. The developed system is found to be a cost-effective system as compared to many other E-Nose systems available, which makes it affordable for the marginal tea farmer. |
