Estimated Spectral Reflectances of Jati Belanda Leaf Using Fourier Transformation

Abstract

This research proposes new method to estimate spectral reflectances of Jati Belanda (Guazuma ulmifolia) leaf based on digital image color. Jati Belanda is a medicinal plant containing antioxidant compounds. Spectral reflectances can be used to estimate chemical compounds but the tools to obtain spectral reflectances is very expensive. To estimate the Jati Belanda leaf reflectance values based on digital image color, we used Fourier Transformation. The estimated spectral reflectances is evaluated using root mean square error (RMSE) and goodness-of-fit coefficient (GFC). According to the experiment, dataset used is 276 leaves of medicinal plants and the best polynomial transformation is 3^rd order with RMSE value 4,83 and GFC value 0.96. The future work could estimate the amount of chemical compounds in Jati Belanda Leaf using estimated spectral reflectances. Estimated spectral reflectances is  an important part of estimated the amount of chemical compounds.