Abstract:
Tea cultivation plays a vital role in global agriculture, but diseases affecting tea leaves
significantly reduce crop yield and quality. Early detection and diagnosis of tea leaf
diseases are crucial for effective management and prevention of economic losses. This
study presents a deep learning-based approach utilizing Convolutional Neural Networks
(CNNs) to identify and classify tea leaf diseases with high accuracy. By leveraging
advanced image processing and feature extraction capabilities of CNNs, the proposed
method automates disease detection from images of tea leaves. The system was trained and
validated on a robust dataset of diseased and healthy leaf samples, achieving promising
results in terms of accuracy, precision, and recall. This approach provides a cost-effective
and scalable solution for tea farmers and agronomists, enabling timely intervention and
sustainable crop management practices. In this research, we leverage deep learning models,
including EfficientNetB4, VGG19, VGG16 and InceptionV3 Model, for the automated
detection of tea leaf disease. The datasets, based on tea leaf disease features contribute to
the comprehensive analysis. The study aims to streamline the diagnosis process by
automating the identification of potential indicators of tea leaf disease, thereby facilitating
early intervention. Comparative analysis of the models reveals varying accuracies, with
our proposed model demonstrating notable performance in tea leaf disease recognition,
achieving an accuracy of 94.94%. These results underscore the potential of deep learning
techniques in enhancing the precision and efficiency of tea leaf disease diagnosis. The
integration of advanced technology to complement existing diagnostic methods, offering a
promising avenue for early tea leaf disease detection.
