Comparison of neural network classifiers for NSCAT sea ice flag

The NASA Scatterometer (NSCAT) is designed to measure wind vectors over ice-free oceans. To prevent contamination of the wind measurements, by the presence of sea ice, algorithms based on neural network technology have been developed to classify ice-free ocean surfaces. Multi-layer perceptron (MLP), radial basis function (RBF) neural networks trained using normalized radar cross section measurements from Ku-band NSCAT Scatterometer are described and compared. Algorithm skill in locating the sea ice edge around Arctic and Antarctic regions is evaluated by comparisons with surface truth (SSMI and SAR images). Classification results show the usefulness of using neural network techniques in flagging ice-free cells in real time and independently of other sensors