CLASSIFICATION OF CORN PLANT DISEASES USING VARIOUS CONVOLUTIONAL NEURAL NETWORK

Abstract

Based on data from the East Java Badan Pusat Statistik (BPS) in 2020, corn production in 2019 decreased by 622,403 tons. The decrease in production was caused by a disease that attacked corn plants identified from the corn leaves' physical appearance. This study aims to obtain an architectural model with good performance between AlexNet, LeNet, and MobileNet in detecting diseases of maize plants. The dataset used in this study came from Kaggle, with 4188 images divided into four disease classes: Common Rust, Gray Leaf Spot, Blight, and Healthy. Agricultural experts from Bantul have confirmed the appearance of each class of corn plant diseases.  The preprocessing process is carried out to prepare the data so that the amount of data for each class is balanced. The image data used in this study totaled 4000 images which were divided into training data and testing data with a ratio of 80:20. Based on the experimental results, it was found that the MobileNet architecture has better performance than AlexNet and LeNet with an accuracy value of 83.37%, average precision of 0.8337, and g-mean of 0.8298. These results have been validated by agricultural experts in Bantul Regency and corn farmers experienced in corn farming.