Martin Rogers MSc BSc (Int)

PhD candidate, Cambridge Coastal Research Unit

I am a PhD candidate in the Cambridge Coastal Research Unit (CCRU). My current research focusses on the application of supervised machine learning tools to automatically detect shoreline position and dynamics in coastal remote sensing imagery.

Biography

Career

• October 2018- PhD candidate. Funded through the Data, Risk and Environmental Analytical Methods (DREAM) Centre for Doctorate Training (funded by the Natural Environment Research Council)
• May 2015- September 2018: National Flood Adviser, National Farmers Union
• May 2015- May 2016: Environment Adviser, National Farmers Union
• October 2011- May 2015: Environment Officer, Environment Agency

Qualifications

• 2014-2016: MSc (Distinction) in River Basin Dynamics and Management with Geographical Informational Systems (GIS), University of Leeds (part-time)
• 2007-2011: BSc Int (First Class) in Physical Geography, University of Leeds. International year of study at School of Earth and Environmental Sciences, University of Queensland, Australia

Awards and scholarships

• Artificial Intelligence for Environmental Risk (AI4ER) affiliated student, August 2019 – present.
• University of Leeds Alumni Bursary, £560, October 2016
• University of Leeds School of Geography Scholarship, £1500, October 2014
• Deans Award for exemplary work, University of Queensland, June 2010
• Deans Award for exemplary work, University of Queensland, December 2009

Research

My research currently focusses on the application of two supervised machine learning tools, convolutional neural networks and support vector machines, to detect the coastal vegetation edge in Landsat-8, Sentinel-2 and Planet 3 and 5 m spatial resolution imagery. My research also uses fieldwork methods, including RTK-GPS data collection and structure from motion, to validate the performance of different computer-based automated methods. I trained a convolutional neural network to produce a python-based publicly available tool, VEdge_Detector, which detects vegetation edges along sandy and shingle beach and dune systems. The tool is available for others to apply to their own area of interest on my Github page.

I have a particular interest in improving our understanding of annual to decadal coastal zone dynamics at the engineering scale (10’s metres to 10’s kilometres) by capitalising on the recent developments in satellite image availability and spatial resolution. I have a broader interest in the application of machine learning tools in the geosciences and was a committee member for the AI4ER seminar series 2019-2020.

Publications

Peer reviewed publications

• Rogers, M. S. J., Bithell, M., Brooks, S. M and Spencer, T. 2021. VEdge_Detector: automated coastal vegetation edge detection using a convolutional neural network. International Journal of remote sensing. http://dx.doi.org/10.1080/01431161.2021.1897185
• Arundell, K.L., Dubuffet, A., Wedell, N., Bojko, J., Rogers, M.S.J and Dunn, A.M., 2019. Podocotyle atomon (Trematoda: Digenea) impacts reproductive behaviour, survival and physiology in Gammarus zaddachi (Amphipoda). Diseases of aquatic organisms, 136(1), pp.51-62. https://doi.org/10.3354/dao03416

Reports

• Rogers, M. and Möller, I. 2019. Suffolk Marine Pioneer Deben. Wave protection provided by salt marshes. A GIS and remote sensing basic analysis approach. Cambridge Coastal Research Unit, Department of Geography, University of Cambridge.
• Stratford, C., House, A., Old, G., Acreman, M., Dueñas-Lopez, M. A., Miller, J., Rogers, M. and Newman, J. 2017. Do trees in UK-relevant river catchments influence fluvial flood peaks? A systematic review. 30pp. Centre for Ecology & Hydrology. Wallingford, UK
• Rogers, M. 2016. Can Structure from Motion determine the efficacy of river re-meandering as a natural flood management measure? Available online: https://www.researchgate.net/profile/Martin_Rogers7

Conference contributions

• Rogers, M. S. J., Bithell, M., Brooks, S. M and Spencer, T. 2021. VEdge_Detector: automated coastal vegetation edge detection using a convolutional neural network. Wavelength2021. University of Sterling, UK [remote].
• Rogers, M. S. J., Bithell, M., Brooks, S. M and Spencer, T. 2021. VEdge_Detector: automated coastal vegetation edge detection using a convolutional neural network. Young Coastal Scientists and Engineers Conference. National Oceanographic Centre, Southampton, UK [remote].
• Rogers, M. 2019. Exploiting satellite technology and machine learning to describe and predict coastal change and hazard regulation. Poster session at Machine Learning and Environmental Science Workshop, British Antarctica Survey, Cambridge.
• Rogers, M. and Möller, I. 2019. Wave protection provided by salt marshes. A GIS and remote sensing basic analysis approach. Suffolk Marine Pioneer Project: Deben Estuary. Ipswich, UK.
• Rogers, M. 2019. Humans on the edge: Exploiting satellite technology and machine learning to describe and predict coastal change and hazard regulation. Oral presentation and poster at the DREAM Symposium, Birmingham.