基于小波和NSCT的图像自适应阈值去噪方法

Adaptive Threshold for Image Denoising Based on Wavelet and Nonsubsampled Contourlet Transform

提出了一种基于小波和非下采样Contourlet变换(NSCT)相结合的图像自适应阈值去噪方法。先用小波估计噪声图像的噪声强弱,再根据噪声的强弱以及NSCT的分解特点及系数所在邻域的特性,给出不同尺度不同方向的自适应阈值。仿真实验结果表明,与小波硬阈值、Contourlet硬阈值和基于非下采样Contourlet硬阈值去噪方法比较,该方法不仅提高了图像的峰值信噪比,减少了Gibbs现象,而且图像视觉效果也明显改善。

A new local adaptive threshold estimation method for image denoising based on the wavelet transform （WT） and Nonsubsampled Contourlet Transform（NSCT） is proposed. The new method uses wavelet to estimate the noise strength of noisy images, then determines the shrinkage threshold according to the strength of noise, the neigilbouring NCST coefficients, the scale of the coefficients and the noise level. Compared with the wavelet hard-thresholding, the contourlet hard-thresholding and the NSCT hard-thresholding denoising method, the proposed method can obviously reduce the Gibbs phenomenon and superiors both in vision and in PSNR（Peak Signal - to- Noise Ratio）.