Harnessing Artificial Intelligence for Sustainable Rice Leaf Disease Classification

Ian Kiew¹Vazeerudeen Abdul Hameed¹Muhammad Ehsan Rana^1*Hrudaya Kumar Tripathy^2*Saurav Mallik^3,4*

• ¹Asia Pacific University of Technology & Innovation, Kuala Lumpur, Malaysia
• ²KIIT University, Bhubaneswar, Odisha, India
• ³Department of Environmental Health, School of Public Health, Harvard University, Boston, United States
• ⁴University of Arizona, Tucson, Arizona, United States

Agriculture encompasses the cultivation of crops, the management of land, and the raising of livestock, all of which are vital for human survival. It plays a crucial role in food security by providing food, raw materials, and livelihoods worldwide. In line with Sustainable Development Goal (SDG) 2, which seeks to eliminate hunger, agriculture is a significant driver of the global economy, contributing 4% to global GDP and up to 25% in some rural areas. Given its close connection to the environment, the sector requires sustainable practices to ensure both long-term food security and ecological stability. Rice, a key crop, serves as a staple food for more than half of the global population, with Asia accounting for 90% of global rice production. Rice is a dietary cornerstone, rich in complex carbohydrates, protein, fiber, iron, manganese, and vitamin B, helping prevent malnutrition. Besides food, rice is used in products like cosmetics. However, rice leaf diseases, like the Hispa, leaf blast, and brown spots, caused by virus, bacteria and fungi, threaten global rice production by reducing crop health, yield, and quality. Advanced technologies, particularly artificial intelligence (AI), offer transformative solutions for rice leaf disease detection and management. AI models, using machine learning and computer vision techniques, can identify disease patterns early, ensuring immediate attention to prevent yield loss. This work aims to develop an AI-driven system for initial detection and classification of rice leaf diseases, enhancing global production of rice and promoting sustainable agriculture. This research utilizes a publicly available dataset consisting of 3,355 labeled images across four categories: Brown Spot, Leaf Blast, Hispa, and Healthy leaves. The dataset diversity ensures a robust foundation for training the AI model for precise classification. The AI model's accuracy and performance metrics ensure reliable disease identification, transforming agriculture with mechanization, cutting-edge practices and supporting SDG 2 by fostering robust agricultural systems and global food security initiatives.