Studies on identification of medicinal plant variety based on NIR spectroscopy using plant leaves

Medicinal plants used very much in herbalism to study the medicinal properties of the plants. The applications of Near-Infrared Spectroscopy (NIRS) have expanded widely in the field of agriculture, plants and various other fields, but the usage for identification of plant variety is still rare. This paper discusses the use of NIRS to distinguish the medicinal plant variety through the spectral signatures of the leaves. Experiments were made on the leaves of five different plants, namely Indian borage, Hibiscus rosa-sinensis, Ocimum tenuiflorum, Solanum trilobatum and Piper betel. Near-infrared (NIR) spectral signatures of the medicinal plant leaves were obtained using XDS™ Optiprobe analyzer from Foss NIR system under reflectance mode. The spectral signatures were collected in two phases, the first set of spectra on the day when the leaves were plucked from the plants and the second set of spectra on dried leaves after three days. Classification model was established for both set of spectral data with suitable preprocessing techniques combined with Principal Component Analysis (PCA). PCA analysis results to discriminate the plant varieties among the dry leaves, which are taken for the experimental purpose. The details of various pre-processing techniques, the analysis and results obtained were presented in this paper. This study shows that NIRS technique can be used to classify for any other medicinal plant varieties.