摘要
分析了自适应范数的变分去噪模型与自适应可信度参数变分去噪模型优缺点,提出了一种同时基于范数p与可信度参数λ全变分自适应图像去噪模型。在该模型中,扩散行为的参数p(x,y)由图像局部梯度决定,可信度参数λ(x,y)大小取决于当前处理位置对应残差图像的特性。实验结果表明,本模型在去除噪声的同时更好地保留图像的细节信息,且峰值信噪比有所提高,取得很好的降噪性能。
Through discussion of the characteristics of the adaptive norm total variation and the adaptive fidelity term total variation, a novel texture preserving adaptive total variational denoising model based on norm p and fidelity term A is pro- posed in this paper. This new total variational model uses the gradient information of each pixel to control the diffusion coef- ficient p （x,y） and adjusts the fidelity A （x, y） according to residual image＇s character of each pixel. Numerical experiments results show that the proposed model can remove the noise while preserving more image details and gain higher peak signal- noise ratio, and has good performance in image denoising.
出处
《重庆邮电大学学报（自然科学版）》
北大核心
2011年第5期621-625,共5页
Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)
基金
四川省教育厅(10ZA13508ZC029)~~
关键词
图像去噪
全变分模型
自适应去噪
可信度参数
image denoising
total variational model
adaptive denoising
fidelity term
