Study of rubber leaf disease using hyperspectral reflectance

Abstract

Leaf disease in rubber leaves causes a significant effect on latex production, especially Pestalotiopsis sp. This disease has caused a massive leaf fall in many plantations in Indonesia. Hyperspectral-based analysis can identify the difference in spectral leaf due to disease. The site location is Sembawa Rubber Research Institute in Sembawa, Palembang, Indonesia. Further, this study also compared other fungi leaf diseases, namely Oidium sp., Collectotrichum sp., and Corrynesspora sp. The leaf spectral reflectance was measured using Ocean Optics USB 2000+ UV-VIS Spectrometer, which measured the spectral response from 350-850 nm. This study aims at (a) defining the spectral signature of rubber leaf infected by leaf fall disease, b) analyzing whether the spectral response of infected leaf can be distinguished, and (c) defining the most useful wavelengths for discriminating spectral responses. The methodology used in this study is spectral response curve, principal components, and partial least square discriminant analyses. As a result, the average spectra indicated differences in color-infected leaves in the chlorophyll-associated wavelengths 490-589, 681–685, 718-752, and 839-875 nm. At the same time, the identification of disease type based on the spectral response found that it is not separable due to high similarity in the spectral curve.