An Intelligent Mobile Application for Detecting Tobacco Leaf Diseases for Candon City, Ilocos Sur

Abstract

Tobacco cultivation, deeply rooted in global agricultural history, holds significant cultural and economic importance, particularly in regions like Candon City, Ilocos Sur, Philippines. Despite its cultural significance, tobacco cultivation faces challenges, including diseases threatening crop yield. Traditional disease detection methods are inadequate, necessitating innovative approaches. Leveraging mobile technology and advanced machine learning techniques, this study developed an intelligent mobile application for tobacco leaf disease detection in Candon City. Employing the Design Thinking framework, the researcher empathized with stakeholders, defined issues, ideated solutions, prototyped, and tested the application. Using an Agile-Waterfall Hybrid Development Methodology, the system was developed iteratively. A dataset comprising 1,400 images was prepared and a CNN-SVM algorithm was employed for disease detection. The model achieved high accuracy and was deployed into the mobile application using TensorFlow.js. The application features mobile compatibility, real-time disease classification, and user-friendly interface. System usability was evaluated, resulting in high scores across metrics with a grand mean of 6.34, indicating strong user satisfaction and usability. This application represents a technological innovation empowering tobacco farmers and enhancing crop sustainability.