An Automatic Detection of Citrus Fruits and Leaves Diseases Using Enhanced Convolutional Neural Network

Accurate and timely detection of diseases present in citrus crops is a crucial task for effective crop management and the prevention of yield loss. Traditional methods of disease detection, such as visual inspection, can be time-consuming and prone to human error. In this paper, we propose a novel approach for automatic and accurate disease detection using convolutional neural networks (CNNs). By investigating the numerous images of infected citrus fruits and leaves, the proposed CNN model has attained prominent recognition and classification accuracy results. The proposed Enhanced-CNN (E-CNN) model is trained using diverse collection of three benchmark datasets such as A Citrus Fruits and Leaves Dataset, Citrus Pest and Disease dataset and Citrus Leaf dataset. Due to careful investigation on layer details and image pre-processing techniques, the proposed E-CNN model secures remarkable performance in citrus fruit and leaf disease detection and type classification. The proposed model achieves significant improvement in disease detection and classification performance by securing the average f1 score 92.06, precision score 95.14, recall score 96.67, recognition accuracy 98% and classification accuracy 99%. These results are comparatively higher than earlier approaches and show more than 6% improvements in disease detection and classification performance. We believe that this unique approach has the potential to significantly improve disease management practices in the citrus industry, helping to improve crop yield and reduce the spread of diseases. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.