Identifikasi Kesehatan Daun Tanaman Padi Menggunakan Klasifikasi Biner Sehat dan Tidak Sehat dengan Algoritma Convolutional Neural Network (CNN) Di Kabupaten Klaten

Abstract

Rice is a vital food crop in Indonesia, where Klaten has become one of the main rice suppliers with a production achievement of 101 thousand tons in 2020. However, the challenge faced is the attack of diseases such as blast, leaf blight, and bacterial wilt which can result in huge losses in yield if not handled effectively. To address this issue, research was conducted using Convolutional Neural Network (CNN), an algorithm commonly used for image processing. In this study, the process involved two main stages namely Feature Extraction and Fully Connected Layer, utilizing a dataset of 2400 images categorized into healthy and unhealthy classes. The results show a very high level of accuracy, with the highest accuracy reaching 0.9653 and validation accuracy reaching 0.8125, as well as low loss with a total of 20 epochs. Through CNN technology, this research makes an important contribution to monitoring the health of rice plants in Klaten Regency, Indonesia, which is expected to help increase productivity and reduce crop losses.