Published Paper Details:

Agriculture is a primary occupation in India, and it loses 35% of its productivity due to plant diseases. Plants play a crucial role in the agriculture sector, exerting a profound influence on a nation's economy and environmental equilibrium. Similar to human health, plants can be affected by diseases caused by pathogens like viruses and bacteria. Effective plant care necessitates diligent observation, accurate disease identification, and appropriate management strategies. Early detection and prevention techniques for plant diseases are highly beneficial to recover plants quickly and maintain their productivity. Currently, the methods employed for identifying plant diseases involve either human-based or lab-based techniques. This research paper presents an artificial intelligence (AI) model that identifies plant diseases, provides an explanation of the detected disease, and suggests appropriate recommendations. The project "AI-based Plant Disease Detection and Recommendation using Deep Learning" is to improve agricultural output by offering an effective method for detecting plant diseases and actionable recommendations. This platform offers two main modules aimed at enhancing productivity and sustainability. Plant Disease Detection Module: Diseases in crops often go unnoticed until they cause significant damage, leading to reduced yields and financial losses. The "AI-based Plant Disease Detection and Recommendation using Deep Learning" platform introduced a plant disease prediction tool that addresses this issue. Farmers can upload images of their crops to the platform, which uses advanced machine learning algorithms to analyze the images and detect potential diseases. Recommendation Module: The Recommendation engine will take the input from the Plant Disease Detection Module, analyze the class of the disease, and suggest the practical, actionable recommendations for prevention and treatment, helping farmers protect their crops and avoid further losses. This platform is developed using Python for backend processing, Flask as the web framework, and HTML, CSS, and JavaScript for the frontend. This system leverages state-of-the-art deep learning techniques to detect and classify plant diseases with high accuracy using plant leaves. InceptionV3 and MobileNetV2 are two distinct deep learning architectures employed for image classification and object detection tasks. InceptionV3 is for its high accuracy, while MobileNetV2 is for its efficiency and suitability for resource-constrained environments like mobile devices. InceptionV3 achieved a training accuracy of 93.88% and validation accuracy of 90.05%, while MobileNetV2 outperformed with a training accuracy of 98.30% and validation accuracy of 94.38%. The dataset includes 70,000 images covering 14 plants, including apple, blueberry, cherry, corn (maize), grape, orange, peach, pepper, potato, raspberry, soybean, squash, strawberry, and tomato, and classified into 38 distinct classes, including both healthy and diseased categories. This project is a highly valuable contribution to farmers and agricultural users for an efficient, user-friendly, and comprehensive tool for plant disease management, thereby encouraging sustainable farming practices.