PlantGraft: Plant identification application with grafting recommendations using artificial intelligence

This project presents the design and development of PlantGraft, a flutter-based mobile application integrated with a machine learning model and backend APIs to support plant grafting practices. The system provides users with plant identification, graft compatibility prediction, care tips, and a community platform for sharing experiences. Firebase Authentication, Firestore, and Storage were used to implement secure sign in, user management and data persistence, while a custom-trained TensorFlow/Keras model, served via FastAPI, delivers compatibility predictions. Agil methodology guided the development across sprints, with deliverables tracked against a structure WBS and Gant Chart. Evaluation was conducted through functional testing, User feedback surveys, and model accuracy checks. Results show PlantGraft enhances knowledge activity for students, hobbies and professionals and demonstrates the feasibility of combining mobile, cloud and AI technologies for sustainable agriculture. The report contributes to both academic understanding and practical innovation by delivering a working artifact, technical analysis, and reflection evaluation.