An Improved Speckle-Reduction Algorithm for SAR Images Based on Anisotropic Diffusion.

Li Gun,Li Cuihua,Zhu Yingpan,Huang Feijiang
DOI: https://doi.org/10.1007/s11042-015-2810-3
IF: 2.577
2015-01-01
Multimedia Tools and Applications
Abstract:This paper mainly studies the algorithm of anisotropic diffusion for speckle noise removal of SAR images. Because the Gauss curvature driven diffusion method is sensitive to the noise and is of low efficiency on suppressing the speckle noise, an improved denoising algorithm is proposed. The new algorithm introduces the difference curvature as the diffusion coefficients of the function, which solves the problem that Gauss curvature driven diffusion is sensitive to the speckle noise, further, Tukey’s biweight function is used to control the curvature diffusion model, which can not only better protect edges, but also automatically control the diffusion. Numerical experiments show that the improved algorithm can preserve the information of textures, edges while inhibiting the speckle of SAR images.
What problem does this paper attempt to address?