Klasifikasi Jenis Daun Jeruk Menggunakan Pengolahan Citra, Linear Discriminant Analysis, dan Algoritma Support Vector Machine

Abstract

Orange a commodity that has an influence on the economy of the community. Indonesia is also the country that exports the most oranges in Southeast Asia, including Thailand and Singapore. There are several types of oranges that have their own unique taste, one of which is orange leaves. Orange leaves function as a cooking spice to eliminate the fishy smell of fish. However, ordinary people still don't know the types of orange leaves because the shape and color are almost the same. This research presents a novel approach to classifying orange leaf types utilizing image processing techniques, Linear Discriminant Analysis (LDA), and Support Vector Machines (SVM). The primary objective is to develop an accurate and efficient system for identifying various orange leaf diseases based on visual characteristics. The proposed methodology involves capturing high-quality images of orange leaves, extracting relevant features using image processing techniques, and employing LDA to reduce dimensionality while preserving discriminative information. Subsequently, SVM is utilized as a classification model to distinguish between different leaf types. Experimental results demonstrate the effectiveness of the proposed approach in accurately classifying orange leaf types, with an accuracy average of 87.11% for LDA and 95.03% for SVM.