Tea Leaf Disease Detection using Deep Learning Convolutional Neural Network Model

The cultivation of tea plants is essential to the Indian economy. By controlling numerous illnesses in the tea plant, the Indian Tea Board helps tea farmers boost tea yield. Many climatic causes and other variables cause these disorders. Effective care and prevention of many diseases depend on early identification. Using a CNN model, a deep learning technique renowned for its superior feature extraction abilities in image analysis, we suggest a technique in this study for detecting tea leaf illness. The classification of sick tea leaves into several groups is examined in the proposed study using a CNN with numerous layers. The classification is done consisting of following categories such as anthracnose, algal leaf, birds-eye spot, brown blight, gray light, healthy, red leaf spot, and white spot. In proposed work, three optimizers were implemented to analyze the performance of the proposed CNN model. The optimizers used are Adam, RMSprop, and Adamax, and the model was run on 420 epochs, and the learning rate was set to 0.0005. Performance metrics like precision, recall, accuracy, and F1-score have been assessed for the model. Using Adam, RMSprop, and Adamax optimizer, the model achieved 83%, 80%, and 73% accuracy rates, respectively. The results show that the proposed model works best with the Adam optimizer. This early detection tool for tea leaves can help with improved treatment planning.