Klasifikasi Jenis Penyakit Tanaman Jagung Berdasarkan Citra Daun dengan Algoritma EfficientNetV2-S

Abstract

Corn is one of the important food commodities in Indonesia, which is susceptible to leaf diseases such as leaf rust, gray leaf spot, and leaf blight. These disease attacks can reduce both productivity and farmers' income. Manual classification carried out by field officers is still constrained by speed and consistency, necessitating an automated classification system based on digital images. This study proposes a corn leaf disease classification model using a Convolutional Neural Network (CNN) algorithm with the EfficientNetV2-S architecture. The dataset used consists of 2,256 corn leaf images from Kaggle, divided into four classes: Leaf Blight, Common Rust, Gray Leaf Spot, and Healthy Leaf. All images were resized to 224×224 pixels, augmented, and then split into 80% training data and 20% validation data. The training process was carried out with a batch size of 32, a learning rate of 0.000005, and 200 epochs. Performance evaluation was conducted using a confusion matrix with accuracy, precision, recall, and F1-score metrics. The best result was obtained at the 150th epoch with 96% accuracy, with no signs of overfitting. These findings indicate that EfficientNetV2-S can serve as an effective model approach for classifying corn leaf diseases.