AI-Driven Crop Disease Prediction and Management System

Abstract

Abstract—The agricultural sector faces critical challenges due to plant diseases, leading to reduced crop yields, economic losses, and food insecurity. Traditional plant disease detection methods are based on manual inspection, which is time consuming, subjec- tive, and prone to errors. This research presents an AI-powered system that utilizes deep learning, specifically Convolutional Neural Networks (CNNs), for efficient disease identification. The model processes plant leaf images to extract key features, classify diseases, and provide real-time predictions. Integrated with a web-based application, the system allows farmers to upload images and receive instant diagnostic feedback and treatment recommendations. By automating the disease detection process, this system improves decision-making, reduces the reliance on ex- perts, and promotes sustainable farming practices. The proposed approach represents a significant advancement in smart farming, improving early disease identification, and reducing excessive use of pesticides. Index Terms—Agricultural Sector, Plant Diseases, Traditional Detection Methods, AI-Powered System, Deep Learning, Convo- lutional Neural Networks (CNNs), Disease Identification, Real- Time Predictions, Web-Based Application, Decision-Making, Sus- tainable Farming, Smart Farming.

Plant phenotyping relevance

植物葉画像から病害を識別する深層学習手法が研究の中心であり、植物の病害状態を画像から推定するフェノタイピング手法に該当します。

Code and data availability

The paper's plant disease image dataset is the publicly available PlantVillage dataset on Kaggle, explicitly named with URL. The authors' trained model and code are only available upon request, so they do not qualify as public assets.