Summary: | Source at https://doi.org/10.1109/IGARSS.2017.8127449 Icebergs can cause a significant danger for shipping, offshore oil exploration, and undersea or subsea pipelines and production facilities. Synthetic aperture radar (SAR) is very valuable tool of detecting and monitoring icebergs in the often dark and cloud-covered polar regions. Detection of small icebergs floating in nonhomegeous sea clutter environments is a challenging task in remote sensing. In this paper, a new methodology is proposed for automatic identification of icebergs in high-resolution polarimetric SAR images acquired during different seasons. This involves adapting the algorithm to sea-ice conditions, and facing challenges when it comes to high iceberg density, meteorological and oceanographic phenomena in the marginal ice zone causing heterogeneity in the background clutter. The algorithm is tested with time series of RADARSAT-2 C-band quad-polarimetric images to detect icebergs in Kongsfjorden (Ny-Ålesund, Svalbard) that have broken off from the glaciers nearby.
|