Banana (Musa spp.) is one of the most important tropical crops globally, yet its production is severely threatened by foliar diseases such as Moko bacterial wilt and Black Sigatoka. Traditional monitoring methods rely on manual inspections, which are labor-intensive, subjective, and often ineffective for early detection. This study introduces AI-BananaMapping, a web-based system that integrates unmanned aerial vehicle (UAV) imagery and YOLOv8-based instance segmentation to enable automatic detection and geolocation of these diseases in banana crops. The system allows users to upload UAV-captured images, process them to segment infected and healthy leaf regions, and visualize results on an interactive geospatial map. With a user-friendly frontend developed in Streamlit and modular backend architecture, AI-BananaMapping offers a scalable and practical tool for precision agriculture. It supports informed decision-making by providing rapid and accurate disease monitoring at scale. Future developments will focus on enhancing model robustness, integrating multi-spectral data, and expanding applicability to other crops with similar disease monitoring challenges.