Seminar - Global analysis of rivers with deep learning

Speaker: Prof. Patrice Carbonneau, Department of Geography, Durham University (UK) - Monday, 17 April – 3:30 PM | Classroom 2M


Much progress has been made over the last decade to build global inventories of freshwater resources. However, existing freshwater inventories are generally produced for a fixed period in time and/or do not discriminate lakes from rivers.

The emergence of deep learning methods and Big Data platforms such as Google Earth Engine offers a potential solution.

However, many obstacles remain in place and the construction of a data processing pipeline that can effectively deliver global scale river classification is not trivial.

This seminar will present a prototype pipeline that combines deep learning with classical remote sensing and image processing methods in order to produce semantic classes for rivers, lakes and gravel bars from Sentinel-2 imagery and at a resolution of 10 meters.

The method is designed for global scale work and we show it’s output for the non-polar globe. The seminar will then consider challenges and opportunities related to the application of deep learning to fluvial remote sensing.