Publication

Abstract : Strawberry diseases have a notable impression on both crop harvest and quality, highlighting the critical need for employing deep learning models to achieve precise disease identification and classification. This study proposes an automated system for identifying and classifying seven common diseases affecting strawberries: Gray Mold, Angular Leaf Spot, Powdery Mildew Fruit, Anthracnose Fruit Rot, Blossom Blight, Leaf Spot, and Powdery Mildew Leaf. We employ deep learning models trained on a dataset consisting of 1450 training images of diseased strawberries. Four models, including a basic CNN and three transfer learning models (VGG16, EfficientNetV2B1, MobileNetV2), were evaluated based on measures such as accuracy, precision, loss, F1 score, recall, coverage, and misclassification rate. The outcomes demonstrate the usefulness of deep learning in accurately classifying these diseases, with Model D (MobileNetV2) achieving the highest classification accuracy of 98.97%. This study emphasizes the strength of MobileNetV2 in achieving a balance between accuracy and efficiency, making it particularly suitable for practical use in agricultural settings. Additionally, the research highlights how transfer learning can enhance disease detection accuracy, even when working with smaller datasets. By utilizing advanced neural networks, this approach offers a reliable and adaptable solution for automating the identification of strawberry diseases, with significant potential for real-time monitoring through mobile devices and other digital agriculture technologies.