作者: E. Rignot , R. Chellappa , P. Dubois
DOI: 10.1109/36.158863
关键词: Covariance 、 Backscatter 、 Synthetic aperture radar 、 Computer science 、 Image processing 、 Radar imaging 、 Fuzzy clustering 、 Radar 、 Remote sensing 、 Cluster analysis 、 Microwave 、 Covariance matrix 、 Speckle pattern 、 Microwave imaging 、 Polarimetry 、 Image segmentation
摘要: A method for unsupervised segmentation of polarimetric synthetic aperture radar (SAR) data into classes homogeneous microwave backscatter characteristics is presented. Classes are selected on the basis a multidimensional fuzzy clustering logarithm parameters composing covariance matrix. The procedure uses both amplitude and phase information, adapted to presence image speckle, does not require an arbitrary weighting different channels; it also provides partitioning each sample used multiple clusters. Given backscatter, entire classified using maximum posteriori classifier. Four-look SAR complex lava flows sea ice acquired by NASA/JPL airborne (AIRSAR) segmented this technique. >