An Efficient Fine-Tuned Convolutional Neural Network Model for Tea Leaf Disease Classification

The Indian population heavily relies on agricultural income, and unfortunately, pests and plant diseases exert a substantial negative impact on it. Tea production plays an essential role, among the oldest and most globally consumed beverages. Diseases in tea plants are quite prevalent and significantly contribute to crop losses. Detecting plant leaf diseases stands as a pivotal concern in agriculture, aiming to enhance crop yields and overall production. Hence, the automated identification of leaf diseases, employing deep learning techniques such as convolutional neural network (CNN), and transfer learning models play a major role in automated feature extraction and classification. To enhance the accuracy of disease recognition, this study proposes a CNN model for identifying tea leaf diseases. The experiment conducted on the collected dataset from Kaggle demonstrates that by utilizing batch sizes of 64 and running for 20 epochs, the tea leaf disease identification model achieves the highest accuracy of 99.86% and a minimal loss of 0.0077. These results validate the effectiveness of this study and hold profound theoretical and practical significance in advancing the field of intelligent agriculture by incorporating the optimal deep learning model for disease identification.