摘要
为了在图像去噪的同时很好地保留细节信息以及边缘信息,本文提出一种结合非局部均值滤波(non-local mean filter,NLMF)的双边滤波(bilateral filter,BF)图像去噪方法。首先利用改进权值函数的NLMF对含噪图像进行预去噪,然后再由得到的图像计算双边滤波的灰度相似性权值并对含噪图像进行最终去噪,同时采用2种快速算法分别实现非局部均值滤波和双边滤波。实验结果表明:与传统非局部均值滤波算法以及双边滤波算法相比,本文方法极大地减少了算法的运算复杂度,具有更好的去噪效果,较少的耗时。因此,本文方法对于图像去噪质量的提升具有一定的实用价值。
In order to fine-tune the image while preserving the details of the information and edge information, an image denoising method based on bilateral filter（BF）combined with non-local means filter（NLMF） is proposed. Firstly, the noise image is pre-denoised by non-local means filter with improved weight function. Then the noise image is denoised by bilateral filter with the pixel intensity similarity weight calculated based on pre-denoised image. Meanwhile, two fast algorithm are used to speed up non-local means filter and bilateral filter separately. Experimental results demonstrate that compared with NLMF and BF, our method significantly reduces the computation complexity of algorithms, and has better denoising performance and lesser time consuming. Therefore, this method has a certain practical value for the improvement of image denoising quality.
出处
《广西师范大学学报（自然科学版）》
CAS
北大核心
2017年第2期32-38,共7页
Journal of Guangxi Normal University:Natural Science Edition
基金
国家自然科学基金(21327007)
广西研究生教育创新计划项目(XYCSZ2017051)
关键词
非局部均值滤波
双边滤波
权值函数
图像去噪
non-local means filtering
bilateral filtering
weight function
image denoising
