Sea Ice Classification of SAR Image Based on Wavelet Transform and Gray Level Co-occurrence Matrix

In order to reduce SAR sea-ice classification results of speckle noise, and we proposed the classification of SAR sea-ice that we can mix the spatial domain and frequency domain texture together according to GLCM and wavelet feature, we carried out the research of the SAR sea-ice classification by using the probe data of the 3rd Shuttle Imaging Radar (SIR-C). The results show that we can extract the wavelet feature and reduce the influence of noise on the basis of remaining the image edge details. In this way we solved the speckled classification´s inaccuracies that the GLCM cannot overcome, and classification accuracy has also been improved.