一种基于Pearson分布类的图像阈值选取方法  

An image thresholding method based on the Pearson distribution

作　　者：吴荣腾[1] 

机构地区：[1]闽江学院计算机科学系,福建福州350108

出　　处：《闽江学院学报》2010年第5期94-97,共4页Journal of Minjiang University

摘　　要：为更准确地模拟数字图像的目标与背景的灰度分布,提出了基于混合Pearson分布类型的图像阈值方法.设计了EM算法计算混合Pearson分布类的待估参数.Pearson分布类能覆盖更大范围的统计分布曲线,因此能更好地匹配未知分布.同时计算量相对比混合广义高斯分布小.An image thresholding method based on the Pearson distribution is presented to fit the probability density functions of the object and background of the digital image. EM algorithm is developed to estimate the parameters of the mixture of Pearson distribution. Pearson distribution can fit a large variety of statistical curves. Moreover,computing complexity is less than the method based on the general Gauss distribution.

关 键 词：阈值 Pearson分布 EM算法 

分 类 号：O211.63[理学—概率论与数理统计]