Ekstrasi Tekstur Citra Menggunakan Local Binary Pattern untuk Identifikasi Penyakit Tanaman Padi dan Anthurium Berbasis Website

Abstract

This research proposes a web-based application to identify plant disease on Paddy and Anthurium automatically based on the leaf image. This research considers seven types of diseases : three Anthurium diseases and four Paddy diseases. For each type of the diseases 100 images are collected, making the total of images data be 700. Local Binary Pattern Variance (LBPV) was used for the extraction of texture and Probabilistic Neural Network (PNN) is used for classification. The result showed that LBPV can be used to identify plant diseases on Paddy and Anthurium. LBPV with operator (8,1) and (16,2) have the highest accuracy of 85.71%.