摘要
针对图像去噪时产生的"阶梯效应"、边缘线条上的光滑性和线状结构不易恢复性,提出基于图像分解与边缘检测的图像去噪方法,首先用图像分解与边缘检测模型将噪声图像分解为结构部分和纹理部分,并提取边缘信息,然后根据边缘指示函数用P-M扩散和相干增强扩散结合的方法对纹理部分去噪,最后将去噪的纹理部分与结构部分组合得到去噪图像。数值试验结果表明,该方法提高了图像去噪的质量,有效避免了扩散中产生的"阶梯效应",较好地保护了边缘信息,恢复其光滑的线状结构。
As for ＂staircasing＂ generated in image denoising, the smoothness and linear structure of edge being not easy to recover, this paper put forward a method of image denoising based on image decomposition and edge detection. Firstly, a noising image is decomposed into structure component and texture component, at the same time the edge information is detected. Then by using denoising method of combining P-M diffusion with coherence enhancing diffusion with the help of edge information, the texture component is denoised. Finally the process is achieved by adding denoised texture component to the cartoon component. Numerical experiments show that the proposed method can improve the qual- ity of image denoising, effectively avoid the spread of ＂staircasing＂, and recover the smoothness and linear structure of edge well.
出处
《咸阳师范学院学报》
2014年第2期22-25,共4页
Journal of Xianyang Normal University
基金
陕西省自然科学基础研究项目(2011JE011)
陕西省教育厅科研基金项目(2013JK0602)
关键词
结构
纹理
图像分解
图像去噪
cartoon
texture
image decomposition
image denoising
