skip to content

AI for the study of Environmental Risks (AI4ER)

UKRI Centre for Doctoral Training

Artificial intelligence approaches to reconstruct changing landscapes with remote sensing

Carbon credits offer a much-needed mechanism to finance the conservation and creation of natural habitats, thereby slowing global warming and protecting biodiversity. Global adoption of carbon credits derived from nature-based projects is hampered in part by a lack of standardised, scalable and transparent techniques to quantify their benefits for society. A fundamental component of such monitoring systems are time-series maps that show how land-cover has changed through time. New methods, exploiting advances in artificial intelligence are needed to reconstruct land-cover time series from remotely sensed observations. 
This project will bring together satellite observations stretching back to the 1980s with novel methods for ground validation, and modern image processing and computational techniques to reveal how landscapes have and continue to change. The work will focus on the ecosystems expected to play a central role in reducing atmospheric greenhouse gas concentrations, including different forest types and peatlands. For the student’s MRes project we propose using multitemporal satellite imagery to assess historical changes in peatland management. The student would build on the work of Hansen et al. (2013) and Vancutsem et al. (2021) to develop models to detect different pixel-level cycles in soil exposure and vegetation regrowth. The consequences for greenhouse gas emissions would be considered for these different classifications. 

Useful skills for this project include: 
-    Knowledge of machine learning and statistical analysis
-    Programming experience with R and python
-    Understanding of remote sensing
-    Knowledge of ecosystem function and global land use change

The successful candidate will join the Application of Artificial Intelligence to the study of Environmental Risks (AI4ER) Centre for Doctoral Training (CDT) based at the University of Cambridge. The AI4ER CDT programme consists of a one-year Master of Research (MRes) course (two terms of formal teaching via lectures, practicals and team challenges plus a three-month research project),  followed by a 3 year PhD project. Both the masters and PhD project will be based on the above project description. 

For further information about the project please contact Prof David Coomes or Dr Thomas Swinfield.