基于B-Spline插值的自适应去噪方法

B-Spline interpolation-based adaptive denoising method

为了提高在高水平噪声下的去噪效果，提出了一种适合对受随机冲击噪声污染图像重建的两步去噪方法。利用图像的全局和局部信息来定位被噪声污染的像素，对被噪声污染像素利用基于B-Spline插值的方法来进行重建，并且根据被污染像素周围像素的污染情况自适应地调整对其的加权因子。实验结果表明：通过B—Spline插值来恢复被噪声污染像素的方法，对污染噪声的水平具有很好的鲁棒性，特别是在噪声水平高于50%的情况下，重建图像的峰值信噪比PSNR比传统方法平均可以提高2～3dB。

A two-stage denoising method for removing random-valued impulse noise under extremely low signal to noise ratios （SNR） was developed to improve the effectiveness of denoising with high noise levels. The first phase identifies pixels which are likely to be corrupted by noise （noise candidates） using the global and local statistical properties of the image. In the second phase, these noise candidates are restored using a B-Spline based interpolation method. During the restoration, an adjustable coefficient is applied to the neighboring pixels adaptively according to the number of corrupted pixels. Experimental results show that the method is robust to The noise level, and outperforms filtering-based denoising algorithms, especially when the noise level is very high.