Medicinal Leaves Classification Using Hyperspectral Imaging

India is notable for its traditional medicinal leaves. These leaves are not just utilised in home cures which involve use of leaves in regular household ingredients, they are utilised as an emergency treatment for a few common ailments like cough, fever, cold, etc. in Ayurveda. These medicinal leaves hold a superior healing power which varies significantly according to their maturity. The practices that were taken to detect the maturity level of leaves using chemical analysis were much more expensive and tedious. To develop a system to classify the medicinal leaves with their maturity, we have used hyperspectral imaging. We have used three medicinal leaf species which are Tulsi, Neem, and Pudina. For capturing hyperspectral images, we have used a specialised hyperspectral camera. These images give more information about leaves. After collecting the required information for each leaf, some ML classifier models like Support Vector Machine (SVM), Random Forest Classifier, and Logistic Regression were implemented with accuracy scores obtained was 99.76%, 66.91%, and 87.77% respectively. Each species has a subclass of Dry and Fresh which will be verified and tested properly with multiple runs along with its maturity level.