Tom Leaf Vision: Real-Time Detection of Tomato Leaf Diseases Using Deep Learning for Early and Late Blight Classification

Tomato cultivation contributes significantly to agricultural production, but it is highly prone to diseases such as Early Blight and Late Blight, which can severely affect crop yield if not identified at an early stage. These diseases spread quickly under favorable environmental conditions and can cause major losses to farmers. Traditional methods of disease identification depend on manual inspection, which is time-consuming, labor-intensive, and often unreliable, especially during the initial stages of infection. As a result, early symptoms are frequently overlooked, leading to reduced productivity. To address this problem, this paper presents TomLeafVision, a deep learning-based system designed for automated detection of tomato leaf diseases. The proposed approach classifies leaf images into three categories: Healthy, Early Blight, and Late Blight using a Convolutional Neural Network (CNN). To improve model performance, input images captured through mobile devices undergo preprocessing steps such as resizing, normalization, and data augmentation. Furthermore, transfer learning using the MobileNetV2 architecture is applied to enhance classification accuracy while reducing training time. The model is trained using the Adam optimizer with categorical cross-entropy as the loss function. Experimental results indicate that the system performs effectively on unseen data and achieves high accuracy. The proposed solution is user-friendly, cost-effective, and suitable for real-time deployment in agricultural environments.