基于自适应中值滤波器和TV修复的椒盐噪声去除

提出一种两步式椒盐噪声去除方法．第一步利用自适应中值滤波器来识别出被噪声污染的候选点；第二步利用一种边缘保持技术（全变差图像修复）来恢复出被噪声污染的点．由于椒盐噪声点的灰度值和原始像素灰度独立，所以在采用恢复技术时不使用噪声点自身的灰度信息．在噪声率不很高的情况下，这种方法可以获得比现有最好方法更高的信噪比；当噪声率高达70％以上时，该方法的信噪比与现有方法非常接近，但是主观视觉效果（例如边缘保持能力）更佳．

Salt-and-pepper noise removal by adaptive median filter and TV inpainting

A two-stage scheme was proposed to remove salt-and-pepper noise in images. First, an adaptive median filter (AMF) was used to identify pixels, which are likely to be contaminated by noise, as noise candidates. Secondly, an edge-preserving technique, called TV inpainting, was used to restore the noise candidates without fidelity to noise data, since salt-and-pepper noise is not related to the "clean" data at all. The proposed scheme can remove salt-and-pepper noise with a higher signal-to-noise ratio (SNR) than the state-of-the-art technique when the noise ratio is not too high. When the noise ratio is as high as 70%, the proposed scheme can obtain a similar SNR with the existing one, but the edges are preserved better in our result.