A new algorithm fusing the fractal interpolation and the enhanced lee filter and its application to the SAR image’s denoising
Heng‐Zhi Cao, Xianchuan Yu, Junlan Zhang · 2008
Against the shortcoming that the Lee filter and enhanced Lee filter algorithms of the High resolution synthetic aperture radar (SAR) images can’t decrease the speckle noise in the heterogeneous area, according to the characteristics that the image of natural scenery has fractal character, we propose a new algorithm, which fuses the fractal interpolation algorithm and the enhanced Lee filter algorithm. In this algorithm, the pixel values in the point-target area of the enhanced Lee filter are replaced by the result of the fractal interpolation. Use ENVISAT-ASAR in Beijing as source data to experiment. Separately intercept urban and suburban areas (512 * 512 pixels) of Beijing SAR image. Experimental result shows that the improved enhanced Lee filter not only maintain the mean-maintaining capacity and the border details maintaining capacity, but also improve the capability of depressing spackle noise. Experimental data quantitative shows that the improvement of the algorithm is effective.