Interpretation of aircraft sea ice microwave data

This paper deals with the Bayesian classifier and its application for a radiometric discrimination of sea ice types. Microwave data used for these studies have been acquired from a dual-frequency passive nadir-looking radiometer which records radiation from earth at wavelengths of 1.6 and 3.2 cm. In addition, a photographic system and IR radiometer operating at
8-14 \mu
m were used in aircraft experiments. Empirical brightness temperature probability distributions for water surface and several ice types for spring (March-April) and autumn (September-October) are given. It is shown that, in order to improve a radiometric discrimination and at the same time to maintain the spatial resolution, it is necessary to use nonlinear low-frequency filtering.