Enhanced Detection and Restoration of Plant Diseases using Deep Learning Approaches

Abstract - Identification and timely management of plant diseases is crucial to enhancing crop yield and ensuring agricultural sustainability. In this paper, we propose a plant disease detection and restoration system utilizing the Xception model, a deep convolutional neural network (CNN) architecture known for its high accuracy in image classification tasks. The system is trained on the Plant Village dataset, comprising 39 categories—38 disease types and one class representing healthy plants. The proposed model effectively identifies and classifies plant diseases, providing a reliable tool for early intervention in agricultural practices. Furthermore, the system incorporates an AI-powered chatbot, integrated with a comprehensive plant health knowledge base, to offer real-time assistance on crop diseases, preventive strategies, and restoration methods. To enhance practical usability, the platform includes a store integration feature that directs users to relevant pesticides and agricultural products based on the diagnosed condition. This end-to-end solution not only facilitates accurate disease detection but also ensures access to timely and effective restoration support, thereby improving decision-making and outcomes in agriculture. Key Words: Plant Disease Detection, Xception Model, Convolutional Neural Network (CNN), Plant Village Dataset, Image Classification, Smart Agriculture, Chatbot, Crop Restoration, Agricultural E-commerce.