Abstract
In recent years, observation of a wide variety in the earth surface can be done by improvement of remote sensing technology. The purpose in this paper is to recognize a drift ice using synthetic aperture rader (SAR) images. The recognition of the drift ice is achieved by using neural networks (NN). The neural network applies two methods, a BP trained neural network and a Self-organizing map. Training data is image features extracted from SAR images. The number of methods of extracting the features are two, Fourier transform and high-order autocorrelation function (HACF). Furthermore, false color is given to a SAR image. Features are extracted from that imge and are recognized by the neural networks.
