Identification of Diseases in Tea Crops using a Computational Convolutional Neural Network Model for Enhanced Crop Production

Background: A major challenge to agricultural productivity in the tea industry is disease. that affects the quantity and quality of tea leaves produced. The extensive development of computational methods for treating diseases has been widely used due to fast and accurate detection. Methods: The proposed method uses sequential Convolutional Neural Network (CNN) computations with many hidden layers to classify diseased and healthy tea leaves into multiple groups. By enhancing feature identification, this structure increases the criteria for accurate disease detection. The data having 5 diseased and one healthy category is obtained from the Kaggle database. After preprocessing the data, it is split into 80:20 ratios for training and testing steps. CNN is constructed using the Keras Sequential API in Jupiter notebook using Anaconda environment. Result: The total accuracy of the ML neural network training for classification was 98.52%. After 50 epochs of training, the model performed well, achieving high accuracy on training and validation datasets. The examination of the confusion matrix showed that several tea leaf diseases may be identified with high accuracy and few misclassifications. In general, the model demonstrated remarkable precision in differentiating between unhealthy and undamaged tea leaves.