Automated Detection and Classification of Tomato Crop Diseases Using Convolutional Neural Networks

Abstract

Tomato plants, a globally significant horticultural crop, are frequently threatened by a range of diseases that compromise yield and quality. Traditional disease detection methods based on manual inspection by experts are often labour-intensive, time-consuming, and susceptible to human error. This study presents a machine learning-based approach that leverages Convolutional Neural Networks (CNNs) to automate the identification and classification of common tomato diseases using leaf images. A comprehensive dataset, including both healthy and diseased leaf images, was collected, pre-processed, and augmented to enhance model performance under various environmental conditions. A custom-designed CNN model was then trained and evaluated using standard metrics such as accuracy, precision, recall, and F1-score. The model demonstrated high classification accuracy and robustness across multiple disease categories including early blight, late blight, and bacterial spot. Furthermore, the system was deployed as a user-friendly web and mobile application interface, allowing real-time diagnosis in the field. This enables farmers especially in resource-constrained settings to identify and respond to infections early, thereby reducing yield losses and limiting excessive pesticide use. The project underscores the potential of AI-driven solutions in modernizing agricultural practices and promoting sustainable crop management. Recommendations are made for future work to improve model adaptability, extend its disease coverage, and integrate environmental sensor data for multimodal analysis.