Geospatial Seminar Series
Upcoming Events:
🚌Accessibility Analysis and Travel Behaviour: Mapping Travel Sustainability for GB Cities
📆Friday, July 31st at 3:00 pm
📍Join: https://teams.microsoft.com/meet/330768498950306?p=jBd7vzMSFVX9t5sSl6 | Meeting ID: 330 768 498 950 306 | Passcode: Bu7NT9r5
Duncan is an Associate Professor in GIS and Visualisation at the Bartlett Centre for Advanced Spatial Analysis (CASA), UCL. His research focuses on urban sustainability, transport planning, spatial analysis, and data visualisation, with a particular interest in how data can help create more liveable and sustainable cities.
This seminar will explore his researching linking public transport accessibility and travel behaviour across Great Britain, highlighting how cities with strong public transport and active travel networks achieve more sustainable outcomes. It also demonstrates how accessibility analysis can help shape future travel patterns by aligning urban development with sustainable transport infrastructure, showcased through the ENHANCE Driving Urban Transitions online mapping platform.
Past Events:
🚲The 100 Cyclists Project and Beyond: Computer Vision for Road Safety Research
📆Friday, 26 June
Road traffic crashes emerge from complex interactions between road users, infrastructure, and surrounding environments. Understanding these interactions remains challenging, particularly when crash data are sparse, incomplete, or missing.
In this seminar, James explored how computer vision and AI methods can support road safety research and monitoring. The talk introduced the 100 Cyclists Project, where panoramic video data is used to capture the real-world experiences of cyclists navigating urban environments and experiencing near misses.
The seminar also discussed recent work using visual language models and street-level imagery to generate road safety ratings, highlighting how advanced AI methods can improve our understanding of road risk and contribute to safer transport systems.
🗺 Robust and Efficient Visual Place Recognition under Missing and Evolving Visual Data
📆Friday, 15 May
An insightful talk by Zichao Zeng, a final-year PhD student at UCL and EPSRC iCASE researcher in collaboration with Ordnance Survey.
Zichao discussed his work using Visual Place Recognition as a key technology behind visual geo-localisation, AR navigation, and a growing range of GeoAI applications. In this session, Zichao explored how large-scale visual data can support urban analytics while addressing real-world challenges such as missing data, occlusions, seasonal changes, and evolving city landscapes.
The talk showcased cutting-edge research on robust visual place representations, cross-domain generalisation, scalable retrieval, and the future of spatial perception in dynamic urban environments.