PSO-SVM Untuk Klasifikasi Daun Cengkeh Berdasarkan Morfologi Bentuk Ciri, Warna dan Tekstur GLCM Permukaan Daun

Suamanda Ika Novichasari (Universitas Ngudi Waluyo)
Yoannes Romando Sipayung

Jurnal Ilmu Komputer
Vol. 1 No. 1 (2018)

Abstract— Of the two types of superior varieties cultivated cloves, clove types of zanzibar is the best kind. However, when not flowering of the three types of clove leaves indistinguishable from the image. This study uses 4 morphological features of shape, 3 color features and 10 most commonly used GLCM features and apply SVM for classification with Particle Swarm Optimization (PSO) optimization method to improve the accuracy of clove plant classification based on leaf surface image. Results of research on the top surface image classification leaf clovers, PSO-SVM method proposed is shown to have a higher accuracy compared with PSO-SVM method than previous research (Novichasari, S.I., 2015) with an accuracy of 90.5% and AUC 0.944. Keywords— Leaf image classification, cloves, shape, color, GLCM, PSO-SVM