A Novel AI-Driven System and Method for Cacao Variety and Disease Classification and Treatment Recommendation Using Image Analysis

Abstract

Cacao diseases significantly impact crop yield and quality, underscoring the need for advanced systems for early detection and treatment. This study presents an AI-driven system for cacao variety and disease classification, along with treatment recommendations, utilizing deep learning and image analysis for sustainable agriculture. The system, trained on an extensive dataset of cacao plant images, achieves 93.7% classification accuracy for diseases in leaves, fruits, and stems. Additionally, the treatment recommendation system demonstrates 88.0% expert-validated relevance, offering optimized solutions for disease management. Developed as a mobile application, the system provides real-time disease detection, personalized treatment guidance, and a user-friendly interface, achieving a System Usability Scale (SUS) score of 85.3, indicating high user acceptance among cacao farmers and agricultural experts. This research contributes to precision agriculture by combining AI technology with practical solutions for cacao farmers. Future work will focus on expanding datasets, refining the mobile interface, and incorporating explainable AI (XAI) models to further enhance transparency and improve adoption in agricultural practices.