Feature extraction of Jabon (Anthocephalus sp) leaf disease using discrete wavelet transform

Abstract

Jabon (Anthocephalus cadamba (Roxb.) Miq) is one type of forest plants that have very rapid growth until the process of the harvest. One inhibitor is a disease that attacks the leaves in the form of spots and blight that can cause death during the growth process of this tree. The purpose of this process is to detect the object of diseases that attack the leaves of jabon at the time in the nursery. Images of affected jabon leaf disease segmented by reducing the RGB color cylinders to separate the disease object from the background. Reduced channel G-R provides information in the form of disease areas contained in the image of Jabon leaf. Furthermore, the characteristics of leaf disease can be detected well using DWT in the 3-level decomposition process with SVM classification results that can separate both classes of spots and blight by 84.672%.