Towards Automated Mapping of Lake Ice Using RADARSAT-2 and Simulated RCM Compact Polarimetric Data
(Sponsor: Canadian Space Agency)
The presence (or absence) of ice cover on lakes during the winter months has an effect on regional weather events such as lake-‐effect snow. Recent work underscores the importance of considering ice cover for improving predictions from numerical weather forecasting models operated by national agencies over lake-rich countries such as Canada. We are developing an automated method to map ice extent (lake-wide fraction) on lakes monitored through visual interpretation of satellite imagery at the Canadian Ice Service. We are evaluating image segmentation algorithms and RADARSAT-2 non-polarimetric/polarimetric and simulated RADARSAT Constellation Mission compact polarimetric SAR imagery to achieve this goal.
Collaborating institutions: Canada Centre for Remote Sensing, Canadian Ice Service, Torsten Geldsetzer Consultant, University of Waterloo (lead)