Identification and classification of medicinal plants of the Indian Himalayan region using Hyperspectral remote sensing and random forest techniques

Abstract

The Indian Himalaya region harbours approximately 1748 plants, which have been prioritized for medicinal usage. To ease the pressure on these plants, the government of India is encouraging the in-situ cultivation of medicinal plants. As a consequence, Saussurea costus, Valeriana jatamansi, and Picrorhiza kurroa are some of the important crops which are being cultivated on large scale owing to their high market demand, conservation value and medicinal properties. Identification of these plants in the field requires taxonomic skills, which is one of the major bottlenecks in the conservation and management of these plants. In this background, a hyperspectral library of the above three medicinal plants has been prepared by collecting its spectral data from Himachal Pradesh and Uttarakhand states of Indian Himalaya. The Random forest (RF) model was implied on the spectral data for the classification of these medicinal plants which resulted in training accuracy of 84.39 % (kappa coefficient = 0.72) and testing accuracy of 85.29% (kappa coefficient = 0.77). This RF classifier has identified Green (555-598 nm), red (605 nm), and NIR (725–840 nm) wavelength regions suitable for discrimination of the above species.