Remote Sensing (Jan 2024)

Sea Ice Detection from RADARSAT-2 Quad-Polarization SAR Imagery Based on Co- and Cross-Polarization Ratio

  • Li Zhao,
  • Tao Xie,
  • William Perrie,
  • Jingsong Yang

DOI
https://doi.org/10.3390/rs16030515
Journal volume & issue
Vol. 16, no. 3
p. 515

Abstract

Read online

Arctic sea ice detection is very important in global climate research, Arctic ecosystem protection, ship navigation and human activities. In this paper, by combining the co-pol ratio (HH/VV) and two kinds of cross-pol ratio (HV/VV, HV/HH), a novel sea ice detection method is proposed based on RADARSAT-2 quad-polarization synthetic aperture radar (SAR) images. Experimental results suggest that the co-pol ratio shows promising capability in sea ice detection at a wide range of incidence angles (25–50°), while the two kinds of cross-pol ratio are more applicable to sea ice detection at small incidence angles (20–35°). When incidence angles exceed 35°, wind conditions have a great effect on the performance of the cross-pol ratio. Our method is validated by comparison with the visual interpretation results. The overall accuracy is 96%, far higher than that of single polarization ratio (PR) parameter-based methods. Our method is suitable for sea ice detection in complex sea ice and wind conditions.

Keywords