1 已有的椒盐噪声密度估计算法 - 中国图象图形学报.doc

中图法分类号：TP301.6 文献标识码：A 文章编号：1006-8961(200 ) - - 论文引用格式： 分块策略实现图像椒盐噪声密度估计 林亚明，李佐勇，林叶郁 (闽江学院计算机科学系 福州 350108) 摘 要：目的 椒盐噪声是造成图像污染的常见因素之一，椒盐噪声密度的估计对椒盐去噪过程中滤波窗口大小的选择具有指导作用。为此提出了一种基于分块策略的椒盐噪声密度估计算法。方法 算法首先对图像按行列等分后形成多个图像子块，统计每个子块中灰度为0或255的像素点个数并排序，然后根据排序后个数差分值函数特征对子块进行筛选，最后将所有候选子块噪声密度估计值的中值作为对整幅图像噪声密度的估计。结果 为验证算法的有效性，选取了两组不同类型的图像进行仿真，与现有椒盐噪声密度估计算法对比噪声密度估计结果。仿真实验结果表明，当图像自身包含较多灰度为0或255的像素点时，新算法的噪声密度估计精度优于现有各种算法，标准差比现有算法小近一个数量级。当图像自身不包含灰度为0或255的像素点时，新算法也能达到现有算法中最优的估计效果。结论 新算法不仅能准确估计不同强度下的噪声密度，而且适用于自身包含灰度为0或255的像素点多的椒盐噪声图像。 关键词 ：椒盐噪声；密度估计；中值滤波；直方图 Estimation of salt and pepper noise in images based on partitioning strategy Lin Yaming, Li Zuoyong, Lin Yeyu Department of Computer Science, Minjiang University, Fuzhou, 350108 Abstract: Objective Salt and pepper noise is one of the common factors causing image contamination. Estimation of salt and pepper noise has a guiding role in determining the size of the filtering window in denoising. For this reason,we propose an algorithm to estimate salt and pepper noise density based on partitioning strategy. Method This algorithm horizontally and vertically splits image into sub-blocks equally, counts sub-block’s pixels whose gray level is 0 or 255 and sorts all sub-blocks, selects candidate sub-blocks according to the characteristic of difference sequence of the sorted pixel numbers, and uses the median of all the candidate sub-block noise density estimations as the estimation of the whole image noise density. Result To evaluate our approach, two different types of images are processed using the presented approach and the noise density estimation results are compared with existing salt-pepper noise density estimation algorithms. The simulation results show that the noise density estimation accuracy of the new algorithm is superior to existing algorithms, the noise density estimation standard deviation of the new algorithm is smaller than existing algorithms by nearly an order of magnitude when an image itself contains man