AI- Powered Ensemble Deep Learning Framework for Plant Disease Detection in Tea Leaves

The growing incidence of plant diseases in the cultivation of tea plants is considered to be an alarming threat to agricultural productivity and economic viability. The early and precise detection of plant diseases is considered to be vital in order to avoid crop damage and ensure quality crop production. However, the traditional method of manual inspection is considered to be time-consuming, subjective, and less reliable, especially in the case of large-scale farming. Keeping this in mind, the current work proposes an ensemble deep learning model using an AI-based approach for the automated detection of tea leaf diseases through image analysis. The proposed model uses multiple CNN models, namely CNN, VGG, and DenseNet, to extract distinct features from the images of the leaves. The predictions are combined using the ensemble method. The system utilizes preprocessing techniques like normalization and data augmentation to enhance generalization under different environmental conditions. The experimental results show that the proposed ensemble model has better accuracy than individual models, and the overall accuracy of the proposed model is 81.3%. In addition, it is further incorporated with a user-friendly interface for real-time disease prediction. The results show the effectiveness of different deep learning models using fusion techniques for intelligent agricultural disease management systems.