Noise Detection Fuzzy (NDF) Filter for Removing Salt and Pepper Noise

摘要

In this paper, we propose a Noise Detection Fuzzy (NDF) filter, to achieve improved filtering performance in terms of effectiveness in removing salt-and-pepper noise while preserving image details. It operates in a moving window where the update value of the central pixel is a function of the median of the pixels in the window. The proposed NDF filter consists of three sequential stages. Firstly, a noise-detection scheme is developed to classify each pixel to be uncorrupted pixel, or otherwise. Secondly, if a pixel is suspected to be noise, it is not used for determining the update value of other pixels. Thus we can prevent noise pixels from distorting the "correct" update value. Thirdly, the fuzzy filter part will then adaptively assign weights to the recorded pixel values to produce the central pixel update value. Experimental results show that our NDF filter outperforms other standard median based techniques.