Geopose-enabled Camera Imagery Interoperability with Geo-AI in Urban Digital Twins
Keywords: GeoPose, TrainingDML-AI, Urban Digital Twin, GeoAI, Camera Imagery Interoperability, Machine Learning
Abstract. This paper presents a GeoPose-enabled pipeline designed to enhance camera imagery and Inertial Navigation System (INS) data interoperability within Urban Digital Twin (UDT) systems. It addresses critical challenges in data synchronization, georeferencing, and integration by leveraging low-cost tools and open standards. The proposed framework captures, processes, and aligns visual and spatial data, converting them into GeoPose and TrainingDML-AI formats to support advanced Geo-AI applications. This methodology enables seamless integration of heterogeneous datasets, facilitating machine learning tasks such as image-based object detection and geospatial analysis. Key contributions include a scalable and cost-effective solution for integrating urban data, ensuring consistency and accessibility across platforms. By advancing the capabilities of UDT systems, this work provides a standardized foundation for real-time decision-making and enhanced urban analytics, promoting smarter and more efficient management of urban spaces, infrastructure, and resources in rapidly evolving smart cities.