We have another exciting Geographic Data Service PhD opportunity based at the University of Liverpool. This project is sponsored by IBM Research as part of a EPSRC Industrial Doctoral Landscape Award (IDLA). The project will be supervised by Alex Singleton, Geographic Data Service, University of Liverpool; Dani Arribas-Bel, IMAGO, University of Liverpool; Anne Jones, IBM Research. Full details of the project and how to apply are below:
Multi-Scale Urban Remote Sensing for the Liverpool City Region: A Foundation Model Approach to Super-Resolution Environmental Mapping
About the PhD Project
This PhD project aims to advance the IBM Research family of Earth Observation foundation models (IBM-NASA Prithvi-EO-2.0 and IBM-ESA Terramind) through a domain-specific adaptation strategy for urban environments. This will fine-tune pre-trained transformer architectures on paired medium-resolution (10-30m) and high-resolution imagery from the Liverpool City Region (LCR). The approach will integrate multi-spectral inputs (RGB, NIR) with temporal attention mechanisms to ensure physically consistent super-resolution across time series.
The project aims to deliver three primary contributions:
- A transfer learning methodology that efficiently adapts global earth observation models to local urban contexts using limited training data
- Achievement of sub-2m effective spatial resolution from 10-30m Landsat and Sentinel-2 imagery, validated against commercial high-resolution benchmarks
- Practical applications demonstrating the framework’s utility for environmental mapping across the Liverpool City Region.
Who are we looking for?
Essential Requirements:
- Strong undergraduate degree (high 2:1 / 1st) in Computer Science, Data Science, Applied Mathematics, Geography or related field
- Demonstrated proficiency in Python programming and deep learning frameworks (PyTorch/TensorFlow)
- Basic knowledge of geospatial data analysis
- Experience with computer vision techniques, particularly image super-resolution or related tasks
- Experience working with large-scale datasets and distributed computing environments
- Excellent problem-solving abilities and strong communication skills for research dissemination
- Ability to work independently within a collaborative research environment
Desirable Skills:
- Masters degree in Computer Science, Data Science, Applied Mathematics, Geography or related field
- Familiarity with geospatial data processing and remote sensing platforms (Google Earth Engine, QGIS)
- Strong mathematical foundations in linear algebra, statistics, and optimization for implementing transformer architectures
- Previous work with foundation models or transfer learning approaches
- Knowledge of urban geography or environmental science
- Experience with multi-spectral imagery analysis
- Interest in applying AI techniques to real-world urban environmental challenges
Apply Now!
Important Note: Applications may close earlier than the stated deadline if a suitable candidate is identified. We strongly encourage interested applicants to submit their materials as early as possible to ensure consideration.
Please email a one‑page statement outlining your interest in the project and your CV to Professor Alex Singleton (alex.singleton@liverpool.ac.uk) no later than 18th July 2025. Use the subject line “PhD Application – Urban Remote Sensing for the Liverpool City Region”. If you are a suitable candidate you will be asked to apply formally to the University of Liverpool and details will be provided.
Funding
The project is an EPSRC Industrial Doctoral Landscape Award (IDLA) Studentship funded in partnership with IBM. These studentships support innovative research in areas of industrial interest, providing postgraduate students with both academic training and industrial experience.
The studentship is fully funded, including academic fees up to the UK rate (international applicants would be required to top-up the difference – https://www.liverpool.ac.uk/study/fees-and-funding/tuition-fees/postgraduate-research/), a stipend (2025-26 rate is £20,780 per year), a £5k research training support budget, and is available to start from October 2025.
The successful candidate will be based at the University of Liverpool and, as part of their project, will spend at least three months working with at IBM Research Europe, Daresbury Laboratory.
Further Reading
This project builds upon our cutting-edge research in the field, including:
- Singleton, A., Arribas-Bel, D., Murray, J., & Fleischmann, M. (2022). Estimating generalised measures of local neighbourhood context from multispectral satellite images using a convolutional neural network. Computers, Environment and Urban Systems, 95, 101802. https://doi.org/10.1016/j.compenvurbsys.2022.101802
- Szwarcman, D, et al. (2024). Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications. https://doi.org/10.48550/arXiv.2412.02732
- Jakubik, Johannes, et al. (2025). TerraMind: Large-Scale Generative Multimodality for Earth Observation. https://doi.org/10.48550/arXiv.2504.11171
