一种有效的自适应均值滤波算法  

Efficient Adaptive Mean Filtering Algorithm

作　　者：李亚珂[1] 赵海武[1] 李国平[1] 滕国伟[1] 王国中[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《电视技术》2013年第23期10-11,49,共3页Video Engineering

摘　　要：结合离散余弦变换(DCT)以及均值滤波算法,提出了一种有效的图像去噪方法。第一步,首先将含噪图像进行分块,并对每一个图像块做DCT变换,将其转换至离散频域;然后通过判断变换块的交流系数,将当前块归类为平坦块或非平坦块。第二步,针对上一步不同类型的块分类,采用不同的均值滤波器进行滤波。经实验表明,该算法相比传统的均值滤波有一定的性能提升。Combined with discrete cosine transform（DCT） and the famous mean filtering, an efficient image denoising method is presented in this paper. Firstly, block based DCT is used to transform the pixels of the noised image into frequency domain,then each block is classified as a fiat or non-fiat one by determining all AC coefficients of the current block. Secondly,mean filter of different types are performed on the determined block. Experimental resuits show that the pnoposed method can prove the performant of mean filtering.

关 键 词：离散余弦变换 均值滤波 统计模型 图像去噪 

分 类 号：TN941.1[电子电信—信号与信息处理]