Detection of Tomato Leaf Infections

Plant diseases pose a significant threat to agriculture by reducing crop yield and quality, often without early signs noticeable to the naked eye. Among these, tomato crops are especially vulnerable to a variety of diseases that can quickly spread and impact production. This study proposes an effective and automated approach to detecting tomato leaf diseases using deep learning, specifically Convolutional Neural Networks (CNNs). The system is designed to classify tomato leaf images into ten categories: healthy, yellow leaf curl virus (YLCV), bacterial spot (BS), early blight (EB), leaf mold (LM), septoria leaf spot (SLS), target spot (TS), two-spotted spider mite spot (TSSMS), mosaic virus (MV), and late blight (LB). A dataset comprising 16,021 labeled tomato leaf images was used for training. The model was trained across three configurations: 10, 20, and 50 epochs, yielding accuracy rates of 64%, 94%, and 97% respectively. The results show a clear improvement in accuracy with extended training, demonstrating the potential of CNN- based models for practical plant disease detection. This technology can play a vital role in assisting farmers with early disease identification, ultimately leading to better crop management and higher agricultural productivity. Key Words: Convolutional Neural Networks, Max Pooling layer, Yellow Leaf Curl Virus, Bacterial Spot, Early Blight, Leaf Mold, Spectorial Leaf Spot, Target Spot, Two spotted Spider Mite Spot, Mosaic Virus, Late Blight 1.