DETEKSI PENYAKIT BERCAK COKLAT, COKLAT SEMPIT DAN HAWAR MELALUI SPEKTRUM WARNA CITRA DIGITAL DAUN PADI MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK

rito cipta; Khoerunisa; Nurul Mega Saraswati Saraswati; Rizki Noor Prasetyono; M. Zidan Alfariki

doi:10.31849/zn.v5i2.13245

rito cipta universitas Peradaban http://orcid.org/0000-0002-9201-3364
Khoerunisa Universitas Peradaban
Nurul Mega Saraswati Saraswati Universitas Peradaban
Rizki Noor Prasetyono Universitas Peradaban
M. Zidan Alfariki Universitas Peradaban

DOI: https://doi.org/10.31849/zn.v5i2.13245

Keywords: : Deteksi Penyakit Tanaman padi, Deep Learning, Convolutional Neural Network (CNN), Asitektur Convolutional Neural Network (CNN)

Abstract

Rice is a prominent food crop commodity and has high potential in the agricultural sector, where rice is a staple food source for Indonesian people. This rice plant certainly has several obstacles, one of which is the presence of rice plant disease attacks through rice leaf spot which can cause crop failure, causing farmers to experience many losses and resulting in poor crop quality, namely empty or empty rice. The long identification process and if the treatment for this disease is very slow will cause the cost of treatment to swell. The use of digital image processing technology in solving problems in this study is to identify rice diseases through digital images based on the morphology of rice leaf spots. One way is by image classification or object classification in the image. The method that can be used in classifying this image is the Convolutional Neural Network (CNN). The accuracy obtained from the Convolutional Neural Network method is based on the 2 types of architecture used, namely the Letnet-5 architecture produces an accuracy of 85% and the Custom architecture produces an accuracy of 90%.