超小波多阈值估计的图像去噪新方法

New Image Denoising Method Based on Contourlet Transform with Multiple Thresholds Estimate

为了能够更好的保留图像的有用信息,更精确的去除图像的噪声,提出了基于超小波多阈值的自适应图像去噪方法.该方法先通过超小波变换对图像进行多尺度,多方向分析,然后采用WindowShrink和BayesShrink相结合的去噪方法,充分利用原始图像和噪声的信息实现了图像的降噪处理.仿真结果表明,文中方法无论峰值信噪比还是去噪图像的效果都优于小波变换.

In order to better conserve useful information in original images and to more accurately remove noise from the image,the paper presents a new image denoising method.It combined the contourlet transform approach with BayesShrink thresholds and WindowShrink thresholds estimating methods.The new image denoise method realizes multi-scale geometric analysis methods that can better approximate nonlinear high dimension function.The simulation results show that the new method is better than the wavelet method and the single threshold estimate method.