改进贝叶斯阈值的Contourlet域图像去噪方法

Contourlet Domain Image Denoising Method with Improved Bayesian Threshold

图像经Contourlet变换后,可以获得更多的方向信息,重构后可以更好地表达图像的轮廓信息。含噪图像经过非下采样Contourlet(NSCT)分解可得到尺度、方向均不同的系数,对这些系数分别进行不同阈值去噪。实验结果表明:利用改进后的BayesShrink阈值对NSCT分解后的各层各方向系数去噪,可获得更好的视觉效果,提高信噪比和边缘保持度。

After the image is transformed by Contourlet,more direction information can be obtained,and the contour information of the image can be better expressed after reconstruction.After non-subsampled Contourlet transform(NSCT)decomposition of noisecontaining images,coefficients with different scales and directions were obtained.These coefficients were denoised by different threshold values.The experimental results show that the improved BayesShrink threshold can be used to denoise the coefficients of each layer and each direction after NSCT decomposition,better visual effects can be obtained,and the signal-to-noise ratio and edge retention can be improved.