SAR sea-ice texture classification using discrete wavelet transform based methods

Texture finds important application in SAR sea ice description and classification. The wavelet transform is an efficient tool for texture analysis because of its multi-resolution nature. This paper presents a study of SAR sea ice classification based on the discrete wavelet transform (DWT). Two kinds of approaches, the traditional DWT based classifier and the tree-structured wavelet packet (TSW) based classifier, have been applied to SAR sea ice textures as well as to the well-known Brodatz textures. The results show that both methods are efficient in interpreting SAR sea ice textures.