Urban area detection in SAR imagery using a new speckle reduction technique and Markov random field texture classification
暂无分享,去创建一个
The strength of synthetic aperture radar (SAR) as a land observation tool resides in S. Gautama et al. (1998) the sensitivity of radar backscatter to the moisture content of terrain media and to the geometrical parameters of the scatterers in the media (i.e., size, shape, roughness and orientation), and (2) an all-weather, day or night imaging capability. However, SAR images are degraded by multiplicative speckle noise due to interference between individual scatterers within one resolution cell. The authors first propose a new speckle reduction method, which preserves edges and doesn't need parameters to be adjusted, based on wavelet decomposition. Comparison with standard speckle filters shows that their filter removes speckle better, while preserving the same amount of detail. Next they use the filter technique in combination with the Markov random field (MRF) texture classification as stated in by S. Gautama et al. (1998) to detect urban areas in the images. The results show that classification results are better when using their proposed filter, compared to the classification results with images filtered with standard speckle filters.
[1] Rama Chellappa,et al. Classification of textures using Gaussian Markov random fields , 1985, IEEE Trans. Acoust. Speech Signal Process..
[2] Giovanni Ramponi,et al. Adaptive image smoothing using fuzzy operators , 1995 .
[3] Stéphane Mallat,et al. A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..
[4] Sidharta Gautama,et al. Markov random fields as a SAR texture descriptor for the delineation of urban zones , 1998 .