An object-based SAR image iceberg detection algorithm applied to the Amundsen Sea

An object-based method for automatic iceberg detection from Advanced Synthetic Aperture Radar (ASAR) images has been developed and applied in the Amundsen Sea, Antarctica. The automatic identification is based on brightness and spatial parameters of the ASAR images at five scale levels, and was verified with manual classification in four areas chosen to represent varying environmental conditions. The presented algorithm works comparatively well with images of the ocean in freezing temperatures and strong wind conditions, common in the Amundsen Sea. The detection rate was 96.2% which corresponds to 93.2% of the icebergs area, for all seasons. The algorithm generated 3.8% errors in the form of 'misses' and 7.0% of 'false alarms', mainly caused by the presence of ice floes. The method was applied on 432 radar images acquired in 2011 under different meteorological, oceanographic and sea ice conditions. As an output a map showing the probability of finding icebergs has been created. It shows that high probability coincides with depth contours and indicates a westward drift of the bergs throughout the whole region. (C) 2016 The Authors. Published by Elsevier Inc.