Edge AI Based Plant Disease Detection System

Abstract

The agricultural sector faces continuous challenges from plant diseases that significantly reduce crop yield and threaten global food security. Rapid and precise disease identification, followed by the correct treatment, is essential to minimize these losses. However, access to expert phytopathological knowledge remains limited for many farmers, particularly in rural and resource-constrained areas. To address this challenge, this paper presents a comprehensive AI-driven decision support system that integrates advanced image-based disease detection with intelligent treatment recommendations. The proposed system is composed of two core modules: a deep learning-based disease detection model and a knowledge-based pesticide recommendation engine. The detection module employs a Convolutional Neural Network (CNN) trained through transfer learning on a large and diverse dataset of plant leaf images, enabling it to accurately classify a wide range of crop diseases across multiple plant species. Once a disease is identified, the system automatically passes the diagnosis to the recommendation engine. This engine accesses a structured knowledge base to provide a curated list of suitable pesticides—both chemical and organic—along with their recommended dosages, safety precautions, environmental considerations, and best practices for application. This holistic “detect-and-remedy” approach empowers farmers with instant, reliable, and actionable insights without requiring expert intervention. By combining artificial intelligence with domain-specific agricultural knowledge, the system enhances decision-making, reduces dependency on specialists, and promotes sustainable crop management practices. The proposed framework demonstrates the potential of AI-based tools to revolutionize modern agriculture by making disease detection and management more accessible, efficient, and intelligent