Abstract:
"Tea crops face threats from pathogens like bacteria, viruses, and fungi, leading to diseases that can harm yield and quality. Prompt identification of these diseases is crucial for effective prevention and maximizing tea production. Current detection methods lack accuracy, hampering disease management efforts. Hence, there's a critical need for innovative approaches for early and precise disease detection in tea plants.
To tackle early disease detection in tea plants, this research utilizes state-of-the-art machine
learning techniques like Convolutional Neural Networks (CNNs) and advanced image processing algorithms. With a comprehensive dataset comprising images of six common tea diseases and their recommended solutions, a predictive model is developed to accurately classify disease instances. Integration of CNNs and image processing ensures precise identification of diseased tea leaves, enabling farmers to implement targeted intervention strategies promptly. Moreover, a novel content-based filtering solution recommendation system is introduced, utilizing established algorithms such as TF-IDF and cosine similarity to provide personalized solutions, thereby enhancing disease management practices.
The proposed methodology was assessed using an image dataset comprising seven distinct disease classes of tea leaves alongside a healthy leaf class. Over 7000 images were utilized for training and validation of the machine learning model. An additional 2000+ images were employed to evaluate the classification performance of both the suggested algorithm and state of-the-art methods. Furthermore, a tailored dataset was implemented for customized solutions. The results revealed that the suggested algorithm achieved an average accuracy of 87% on the test data, effectively diagnosing and categorizing diseases impacting tea leaves with the solution. "
