SAR Polarimetry for Sea Ice Monitoring
Alfred Wegener Institute for Polar and Marine Res.,
(2) Technical University of Denmark, Oersteds Plads, Bygning 348, DK-2800 Kgs. Lyngby, Denmark
Polarimetric SAR imaging has shown a promising potential for the classification of sea ice types. However, from the last decade only a few studies are available, each of them focussing on one particular SAR sensor (mostly airborne) and certain polar ocean regions. In 2001, a project was initiated by the European Space Agency, which deals with the application of polarimetric radar in various geo- and biophysical studies and data surveys. The goal is to quantify the improvement of information gathering gained by utilizing polarimetric systems instead of the single-polarization SAR instruments, which are part of recent remote sensing satellite missions. In the project, polarimetric classification methods had to be investigated with regard to their accuracy and their robustness in operational use. To this end, different already existing data sets had to be combined For our sea ice study, test sites were selected in different ice regimes (Greenland Sea, Baltic Sea, and Beaufort Sea). For these test sites, at least a minimum of complementary data about the state of the ice and about the environmental conditions are available. The polarimetric data were acquired by the Danish EMISAR and the US AIRSAR which both are airborne systems. In our study, we focus on the use of C- and L-band data. Because of its robustness, a knowledge-based classification is regarded optimal. The potential of a number of polarimetric parameters for ice type discrimination is assessed. On the basis of this assessment, a classification scheme is devised and tested. In our presentation, we will compare the classification accuracies of polarimetric and single-polarization SAR modes for the different ice regimes. The merits of a polarimetric SAR for the discrimination of sea ice types will be discussed.