Drone-Based Maritime Geospatial Intelligence: Automated Vessel Detection, Positioning and Characterization

Unmanned Aerial Systems are increasingly used for inspection and mapping in maritime environments, but their integration into operational geospatial workflows remains challenging. This contribution presents a drone-based image analysis framework for transforming aerial imagery or video into actionable maritime geospatial information. The approach combines photogrammetric reconstruction, computer vision and artificial intelligence to detect, locate and characterize vessels in port and coastal environments.

In a port context, drone imagery is processed through aerial triangulation and bundle block adjustment to generate georeferenced 3D point clouds. Vessels moored along quay walls are extracted from these point clouds and positioned relative to bollards along the quay wall, enabling the derivation of operational information such as berth occupancy, vessel extent and relative position along the quay. In a complementary maritime surveillance context, drone imagery can be used to detect vessels in open-water environments and estimate their position through direct georeferencing. Image observations are projected onto a sea-surface model using calibrated camera geometry together with GNSS/IMU-derived position and orientation information, avoiding the need for a full 3D photogrammetric reconstruction. Depending on payload and mission setup, additional information such as vessel dimensions, identity markings and inspection-relevant visual features can also be extracted.

The framework has been explored through case studies covering port operations and coastal/offshore surveillance. Results demonstrate centimetre-level geometric accuracy for photogrammetric reconstruction in port environments and vessel localization accuracies on the order of tens of metres in open-water surveillance scenarios. The work also highlights practical constraints for operational deployment, including flight geometry, image quality, camera calibration, payload configuration, automation level and integration with AIS-based vessel information and port digital twin environments.

By combining photogrammetric reconstruction, direct georeferencing, AI-based image interpretation and geospatial analysis, the framework demonstrates the potential of drones as operational decision-support tools for maritime stakeholders.