- Course: Centre for Satellite Data in Environmental Science (SENSE CDT)
- PhD title: Developing a new methodology for assessing future food production
- Nationality: British
- LinkedIn: https://www.linkedin.com/in/samuel-bancroft/
Samuel Bancroft is a postgraduate researcher on the Centre for Satellite Data in Environmental Science programme (SENSE CDT). He is developing a new methodology for assessing future food production based on machine learning, remote sensing, and crop models. Part of the Climate Impacts group, Samuel is supervised by Professor Andy Challinor from the School of Earth and Environment; he is co-supervised by Professors Netta Cohen and Anthony Cohn from the School of Computing.
Being part of a Centre for Doctoral Training (CDT) involves working together as part of student cohort, and includes a specialised training programme during the first year.
Samuel said: “CDT programmes are typically funded in order to address a key area for UK industry. This is why SENSE stood out to me, as it represented a funding commitment to train the next generation of Earth Observation scientists.”
Assessing future food production
Samuel uses satellite imagery and machine learning to analyse crop data on a highly detailed scale. He uses the data to monitor agricultural land worldwide that can give insight into future food production.
“My project is about assessing future food production, using satellite imagery and machine learning,” said Samuel. “A big part of that is identifying the location of different crop species. Earth Observation has its history firmly rooted in agricultural monitoring - with the recent ESA Sentinel satellites providing excellent sources of data to further develop this field.”
He continued: “This, in conjunction with the take-off of deep learning means that many exciting and tangible questions can be answered in ways that used to not be possible. It's a very exciting field that is progressing rapidly - becoming increasingly crucial in monitoring climate change.”
It's a very exciting field that is progressing rapidly - becoming increasingly crucial in monitoring climate change.
“Machine learning in general, not just my research, is constrained by a dependence on labelled data,” Samuel explained. “We don't always have the best ground truth data about what crops are growing where, and this imposes limitations in our models.
“I hope to develop, through my research, methods that improve the ability to classify crops and adapt to unseen areas and time periods (improve generalisability). By doing this, we can generate crop type maps at large scales and use them to improve food security in a future impacted by climate change.”
We can generate crop type maps at large scales and use them to improve food security in a future impacted by climate change.
Find out more about Professor Andy Challinor’s research on modelling climate impacts to strengthen global food security.
For more details about studying for a research degree at Leeds, visit our research degrees pages.