Automatic Identification of Tomato Leaf Conditions Based on OpenCV and Convolutional Neural Networks

Abstract

Tomato leaf diseases are a major cause of yield loss, particularly in rural areas with limited access to agricultural experts and diagnostic facilities. This study proposes an artificial intelligence–based system for identifying tomato leaf conditions using a Convolutional Neural Network (CNN) integrated with OpenCV. The system classifies tomato leaves into nine categories, including one healthy class and eight disease classes, based on digital images. The dataset was divided into training, validation, and testing sets, with pre-processing steps including resizing, normalization, and data augmentation. The CNN model was trained using the Adam optimizer and categorical cross-entropy loss. Experimental results show that the model achieved approximately 90% accuracy, with average precision, recall, and F1-score values above 0.88, indicating strong classification performance and good generalization ability. The OpenCV-based implementation enables real-time detection via a camera with an average prediction time of less than one second per image. These findings demonstrate that integrating CNN with OpenCV provides a practical and efficient solution for early tomato leaf disease detection and supports decision-making in technology-driven agriculture.