SMART AGRICULTURE THROUGH CONVOLUTIONAL NEURAL NETWORKS FOR PLANT DISEASE CLASSIFICATION

Abstract [English]

The Plant Disease Classifier is an AI-powered system designed to transform modern agriculture by enabling accurate and timely identification of plant diseases through image analysis. Utilizing advanced machine learning techniques, particularly convolutional neural networks (CNNs), the system classifies diseases from images of plant leaves, offering real-time diagnostic feedback to assist farmers and agricultural experts in taking proactive measures. This early detection capability is crucial for minimizing crop losses, enhancing yield, and promoting food security.
The project methodology involves the collection and preprocessing of a comprehensive dataset comprising both healthy and diseased plant images, followed by the training and evaluation of a deep learning model using performance metrics such as accuracy, precision, and recall. The final model is deployed via an accessible mobile or web application, making disease diagnosis practical and scalable.
The classifier is capable of detecting a broad spectrum of plant diseases—including bacterial, fungal, and viral infections—while incorporating advanced image processing techniques to improve input quality and model performance. Additionally, the study explores existing literature, outlines current challenges in plant disease detection, and suggests future enhancements such as IoT integration for real-time monitoring and automated health assessments.
By bridging the gap between traditional inspection methods and precision agriculture, the proposed AI solution represents a significant advancement toward smarter, more sustainable farming practices.