一种快速有效的图像脉冲噪声滤除方法

A Fast and Effective Algorithm of Pulse Noise Filtering for Imaging Data

针对图像数据中脉冲噪声的特点,该文提出了一种快速有效的滤除方法。这种方法通过寻找局部极值快速确定噪声点,并对图像的所有像素做分类标记,处理的过程中只考虑标记为噪声的点,采用邻域内非噪声点的均值作为滤波输出。这样处理不会影响图像中未受噪声污染的像素,而且减少了相应的计算量,经测试,该算法的滤波性能和计算速度明显好于常用的中值滤波方法,可以快速有效地滤除脉冲噪声,具有很好的实用价值。

According to the features of the pulse noise in digital image,a fast and effective algorithm of pulse noise fil-tering is proposed in this article.Using this method,the noisy point is determined quickly by looking for the local ex-tremum and the classified markers are made for all pixels of the image.Only the marked noise points are taken account into the processing.The averaging values of non-noisy points in the neighborhood are taken as the output of the filter.By using this algorithm,the uncorrupted pixels would not be affected and the computation was much simple.In compar-ison with the medians filtering algorithm used in image processing frequently,the filtering performance and the comput-ing speed of the proposed algorithm are better than the former.It is useful in the preprocessing of images to reject the pulse noises effectively.