Abstract

A combined supervised classification and unsupervised classification method was used to distinguish ice ridges from the surrounding ice in the Ross Sea based on radar backscatter values. However, brash ice between ice floes also has backscatter values similar to ice ridges. To discriminate ice ridges from brash ice, an object delineation method is then used to-depict ice floes as objects. The high-backscattering objects within ice floes can further be uniquely identified as ice ridge pixels. Finally, the ridge concentration (0.38-0.42%) was extracted based on image processing of an ERS-2 SAR image acquired on November 3, 1997 and the statistics of field-measured ridge heights (1.14-1.26 m). The ridge concentration (0.63-1.80%) was also derived based on ridge statistics and measured ridge heights and occurrence frequencies (0.815-2.125 ridges/km). Statistics of other surface features were also derived based on image processing.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call