Klasifikasi Daun Tomat Sehat dan Terserang Penyakit Menggunakan Metode Convolutional Neural Network (CNN)

Abstract

Tomato plants (Solanum lycopersicum) are an important agricultural commodity that has high economic value, both domestically and globally. . According to data from FAO, world tomato production will reach more than 180 million tons in 2023, with the main producing countries being China, India and the United States. However, despite increasing tomato production, the threat of plant diseases remains a major problem that impacts crop yields and product quality. To deal with this problem, we classified diseases on tomato leaves. Tomato leaves are one of the factors used to see whether tomato plants are good or not. In this study we used the Convolutional Neural Networks (CNN) method which has proven to be effective in classifying plant disease images. This research compares three CNN models, namely the deep model, the wide model, and the combined wide & deep model. Test results show that the deep model provides the best performance, with the highest accuracy of 100%. The average macro and weighted values ​​for all classes in the deep model show a precision level of 0.99, recall 0.99, and F1-score 0.99, which reflects the consistency and accuracy of the model in classifying each class accurately.