摘要
主成分分析(PCA)是一种将多个变量通过线性变换选出较少个数重要变量的一种多元统计方法。在图像去噪中,由于图像的局部相似性,提出一种新的有效的去除噪声的算法。通过块匹配法寻找出相似块作为训练样本,利用主成分分析提取信号的主要特征,然后根据统计理论中最小均方误差方法构造线性自适应阈值方程,对含噪图像的每一块进行自适应阈值去噪。实验结果表明,该方法能有效去除图像的高斯白噪声,并同时能很好的保持边缘等的细节信息。
Principal component analysis （PCA） is a multivariate statistical method which selects a few important variables through a linear transformation of Multiple variables. In image denoising, because of the local similarity of images, a new and effective noise removal algorithm is put forwarded. The similar blocks are found out as training samples by block matching algorithm and the main signal feature extraction is extracted by PCA, and then, an adaptive threshold is used to each denoised block to remove noise. The experimental results show that the method can effectively remove the image of Gauss white noise, and at the same time, can be very good to keep the edge detail information.
出处
《红外技术》
CSCD
北大核心
2014年第4期311-314,319,共5页
Infrared Technology
关键词
主成分分析
块匹配
自适应阈值
图像去噪
principal component analysis, block match, adaptive threshold, image denoising
