Plant Care and Disease Detection Using Pattern Recognition

Abstract

The well-being of plants is paramount for maintaining a healthy environment and sustaining life on Earth. In this study, we developed a mobile application called "Plant Care and Disease Detection Using Pattern Recognition" to address these challenges. The application uses a deep learning model based on Convolutional Neural Networks (CNN) to analyze plant leaf images and determine the presence of diseases with 97% accuracy and an F1-score of 98%. The app connects to a pre-trained model with 37 classes, including 13 different plant diseases and healthy plants, totaling 125,319 images. Users can scan a plant leaf with their mobile device's camera, and the model provides a diagnosis or confirms that the plant is healthy. The app schedules watering and fertilizing tasks based on specific plant needs, with notifications to remind users of care times. Additionally, the app uses Firebase for authentication and Firestore for database management, allowing users to sign up, create profiles, and recover passwords if forgotten. This application improves plant care by offering a reliable solution for disease detection and care scheduling. The broader implications suggest that this technology can support sustainable agricultural practices and contribute to global environmental efforts by reducing the need for chemical treatments through early detection. Ultimately, the "Plant Care and Disease Detection Using Pattern Recognition" app showcases how deep learning can transform plant care, fostering a healthier plant ecosystem and benefiting individual gardeners, the agricultural sector, and the broader environment.