Investigating the potential of different polarimetric features based on dual polarimetric TerraSAR-X data for automated sea ice classification

Abstract

In this work, we examine the performance of an automated sea ice classification algorithm based on dual polarimetric TerraSAR-X data. Polarimetric features are extracted from HHVV dualpol stripmap images. In a second step, the feature vectors are fed into an artificial neural network to classify each pixel into an ice type. The first part of our analysis addresses the predictive value of different subsets of features for our classification process (by means of measuring mutual information). Different neural network configurations are then explored for optimal classification performance. The results on a TerraSAR-X dataset indicate a high reliability of a trained dual polarimetric classifier. Performance speed and accuracy promise applicability for near real time operational use.