Human-Centred GeoAI: automating national mapping updates through interactive change detection
Ordnance Survey maintains a national vector map (Ordnance Survey National Geographic Database) that requires frequent updates to reflect real-world changes. The current update process, largely manual, faces challenges in automation, scalability, timeliness, and cost-efficiency. Introducing automation supported by human-centred design could transform this workflow, combining AI efficiency with expert oversight.
The Challenge Detecting and acting on changes between two versions of aerial imagery is complex. The solution we are envisaging must not only identify OS-relevant topographic changes (e.g. buildings, street lights, field boundaries) but also integrate human validation and interaction to ensure trust and usability. Designing intuitive interfaces and workflows for map editors is key.
Researchers: Ed Daniels, Russell Summers, Sophie Maw and Ikechukwu Ofodile