图像中的高斯和脉冲混合噪声处理

通常情况下,受脉冲噪声污染的图像采用中值滤波法复原,受高斯噪声污染的图像采用均值滤波法复原。为了去除两者的混合噪声,Lee和Kassam提出了一种改进的均值滤波算法ModifiedTrimmedMean(MTM),但是MTM的使用受到了阈值的限制。为了在滤除退化图像中混合噪声的同时能更好地保护图像细节,我们详细分析了MTM滤波的特点,在对MTM进行改进的同时,提出了一种改进的自适应中值滤波算法(统计滤波)。该算法无需噪声的先验知识,利用VisualC 自动搜索阈值来实现图像的最佳复原。利用两种客观标准进行评价,实践证明新方法的处理结果优于传统的MTM方法。

Image processing of Gaussian noise and impulse noise

Usually, median filter is used to process the image, which is polluted by impulse noise, and trimmed mean filter is used to process the image, which is polluted by Gaussiuan noise. In order to process the mixed noise, Lee and Kassam proposed an improved algorithm of mean filter, but a threshold that consumedly affects the effect of MTM. In order to preserve more details while erasing mixed noise, modified trimmed mean (MTM) was studied in detail, and we proposed an improved method of median(SF). This algorithm doesn't need any information of noise. To recover the image best, Visual C was used to search threshold automatically. Two external criterions was used to appraise it, the new method shows better performance than the MTM filter in real application.