Development of hyperspectral indices for anti-cancerous Taxol content estimation in the Himalayan region

Abstract

Monitoring and management of rare and economically important species in the highly complex terrain are challenging and thus need advanced technological development. In this study, the hyperspectral radiometer data of Taxus wallichiana were acquired at highly complex terrain of the Pindari region of the Himalaya and processed by using several sophisticated algorithms to deduce Taxol content in the plants. The spectroradiometer data were denoised through three different types of smoothing filters such as Average Mean, Savitzky Golay, and Fast Fourier Transform (FFT) followed by feature selection for allocation of best bands for Taxol content estimation. The results showed that the Average Mean filter in combination with feature selection performed best for Taxol spectral indices generation, model development, and Taxol content prediction. The best model showed a correlation of 0.719 with a relative root mean square error (RMSEr) value of 0.678 for Taxol content prediction. © 2022 Informa UK Limited, trading as Taylor & Francis Group.