使用邊緣保留與相鄰視窗變化調整法於受干擾影像中移除椒鹽雜訊

數位影像經常會受到脈衝雜訊干擾，導致影像口質衰退，甚至變得模糊不清，導致無法辨識原始的影像資訊;因此如何移除受干擾影像中的雜訊，是一項非常重要的研究工作。本文提出新的方法來移除受到雜訊干擾影像中的椒鹽雜訊:我們採用像素間變動的一致性來有效的偵測受雜訊干擾的像素，然後施予均值慮波，重建原始清晰影像的像素值。此外，本文也考慮相鄰像素變動必須保持關聯性的方式，使得邊緣變動方向，在相鄰視窗間具有一致性，提高邊緣影像像素重建的正確性。實驗結果證實:本文提出來的方法確實可以在中、低雜訊密度的環境下，有效的移除受干擾影像中的椒鹽雜訊。

The quality of digital image can be seriously deteriorated by the corruption of impulse noise. How to efficiently remove this noise for a corrupted image is an important research task. This paper proposes a new method to remove the salt-and-pepper noise by keeping the minimum variation in neighbor pixels when mean filtering is performed. This enables the noise corrupted image to be efficiently restored. In addition, the motion direction is maintained to be consistent for the edge pixels of an object, yielding the quality of restored image being improved. Experimental results show that the proposed method can efficiently remove the salt-and-pepper noise for a corrupted image in low to middle noise densities.