一种最近邻像素值的极大概率滤波算法  

A great probability filter algorithm based nearest-neighbor pixels' gray

作　　者：梁立东[1] 叶邦彦[1] 

机构地区：[1]华南理工大学机械与汽车工程学院,广东广州510640

出　　处：《电子设计工程》2014年第7期181-183,187,共4页Electronic Design Engineering

基　　金：国家自然科学基金资助项目(50875089)

摘　　要：图像增强是图像处理中重要的过程,其目的是消除噪声和凸显感兴趣特征。针对当前经典中值滤波算法不足,提出一种最近邻像素值的极大概率滤波算法(GPF),该算法将掩膜窗覆盖下的所有像素值视为一个像素值集,并定义评价各个元素对整个集合的影响度关系,同时提供计算这种影响度的方法,确定选值准则。最后在VC++6平台上实现该算法并与当前存在的中值滤波算法相比较。实验证明,关于最近邻像素值的极大概率滤波算法能够有效消除噪声和保留细节。Image enhancing is an important process when image processing ,which aims at denoising and highlighting interesting parts in an image.Directing at the disadvantage of traditional median filter ,presents the Great Probability Filter Algorithm （GPF） based nearest-neighbor pixels＇ gray value.The Algorithm treats the pixels＇ value as a value set,and defines a Influence Degree between an element and the set,and determine the GPF Criterion.And finally realize the GPF algorithm in VC＋＋6 and compare it with traditional median filter algorithm .Experiments show that the GPF algorithm can denoise image effectively and attain detail in image.

关 键 词：消除噪声 滤波算法 极大概率 影响度 中值滤波 

分 类 号：TN919.8[电子电信—通信与信息系统]