SAR Image classification of first-year ice types, Bay d'Espoir, Newfoundland and Labrador

Brett, Paul (2004) SAR Image classification of first-year ice types, Bay d'Espoir, Newfoundland and Labrador. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (17MB)

Abstract

Sea ice during spring melt and breakup can prove dangerous to infrastructure located in the coastal zone. Industries such as aquaculture, which houses much of its infrastructure in the near shore environment, are at the mercy of sea ice motion. Knowledge of the condition of ice during the melt season could allow such users of the coastal zone to plan against potential damage due to the premature breakup of the ice coverage. -- Traditional methods of sea ice detection and classification have been limited to the open ocean and dedicated to problems associated with navigation. Their primary concern is in the identification of first-year ice, multi-year ice and open water. -- The use of second order texture measures, along with a new approach to histogram characterization, and neural network classification have allowed the classification of a fine beam mode RADARSAT image to map five sub classes of first-year sea ice – brash, puddle, flooded, rotten, snow covered, – and open water. Classification accuracies achieved are on the order of 60% with user's accuracies for several of the ice types approaching 100%.

Actions (login required)

View Item