自适应迭代维纳滤波算法
|
摘要
针对经典维纳滤波去除图像高斯噪声后不能很好保持图像边缘和平滑性的问题,提出一种自适应迭代维纳滤波算法。首先,依靠像素灰度值的二阶差分分量估计图像的噪声方差;然后,以3×3大小的模板为起始模板,对图像进行迭代滤波且每次迭代都增加模板大小;最后,每次滤波前后所有像素灰度值变化大小的均值是否小于阈值为停止迭代的判断条件,判断条件成立结束滤波。不同改进算法的实验结果对比,该改进算法不仅保持较高图像峰值信噪比而且在时间复杂度上要小于小波域维纳滤波器,并且在保持图像边缘和平滑效果方面要较好于小波域维纳滤波。
The smoothness and edge of the image cannot be well kept, this paper proposes an adaptive iterative Wiener filtering algorithm according to the classical Wiener filtering to remove image noise. Firstly, the second-order differential components of pixel gray value estimates the noise variance of image: then, the starting filter template size is the 3×3 template, the template size increases after each iteration filter. The last, the judgment conditional of stop iteration that the average of all pixels ' gray value change is less than the threshold, if the judgment condition is established, adaptive iterative wiener filter is stopped. Through comparison of experimental results of different improved algorithms, the improved algorithm not only maintains a high PSNR and the time complexity is less than the wavelet domain wiener filter, and the image edge and smoothing effect is better than Wiener filtering in wavelet domain.
The smoothness and edge of the image cannot be well kept, this paper proposes an adaptive iterative Wiener filtering algorithm according to the classical Wiener filtering to remove image noise. Firstly, the second-order differential components of pixel gray value estimates the noise variance of image: then, the starting filter template size is the 3×3 template, the template size increases after each iteration filter. The last, the judgment conditional of stop iteration that the average of all pixels ' gray value change is less than the threshold, if the judgment condition is established, adaptive iterative wiener filter is stopped. Through comparison of experimental results of different improved algorithms, the improved algorithm not only maintains a high PSNR and the time complexity is less than the wavelet domain wiener filter, and the image edge and smoothing effect is better than Wiener filtering in wavelet domain.
引文
[1]LEE J S. Digital image enhancement and noise filtering by use of local statistics[J]. IEEE Tranactions on Patten Analysis and Machine Intelligenll. 1980,PAM(1-2)(2):165-168.
[2] KUAN D T,SAWCHUK A A, STRAND T C, et al. Adaptive noise smoothing filter for images with signal-dependent noise[J].IEEE Transactions on Pattern Analysis&Machine Intelligence, 1985,7(2):165.
[3]JIN F, FIEGUTH P, WINGER L, et al. Adaptive wiener filtering of noisy images and image sequences[C]//International C.onference on Image Processing, Spain Barcelona,2003.
[4]张小波.基于维纳滤波的图像去噪算法研究[D].西安:西安电子科技大学,2014.
[5]白俊奇,赵春光,王寿峰,等.红外图像中的自适应维纳滤波噪声抑制技术[J].光电工程,2011,38(11):79-85.
[6]李东兵,李国平,滕国伟,等.一种新的自适应维纳滤波方法[J].上海大学学报:自然科学版,2012, 18(6):555-560.
[7]刘红亮,陈维义,许中胜.一种新的小波域维纳滤波图像去噪算法[J].海军工程大学学报,2015, 27(6):63-67.
[8]ZHANG X, FENG X. Multiple-step local Wiener filter with proper stopping in wavelet domain[J]. Journal of Visual Communication&Image Representation,2014,25(2):254-262.
[9]颜军,郝雪,王兆国.Wiener滤波器在小波变换域的应用[J].吉林大学学报,2006(s2):67-69.
[10] LAHMIRI S. An iterative denoising system based on Wiener filtering with application to biomedical images[J]. Optics&Laser Technology,2017, 90:128-132.
[11] IMMERKETR J. Fast noise variance estimation[J]. Computer Vision&Image Understanding, 1996, 64(2):300-302.
[2] KUAN D T,SAWCHUK A A, STRAND T C, et al. Adaptive noise smoothing filter for images with signal-dependent noise[J].IEEE Transactions on Pattern Analysis&Machine Intelligence, 1985,7(2):165.
[3]JIN F, FIEGUTH P, WINGER L, et al. Adaptive wiener filtering of noisy images and image sequences[C]//International C.onference on Image Processing, Spain Barcelona,2003.
[4]张小波.基于维纳滤波的图像去噪算法研究[D].西安:西安电子科技大学,2014.
[5]白俊奇,赵春光,王寿峰,等.红外图像中的自适应维纳滤波噪声抑制技术[J].光电工程,2011,38(11):79-85.
[6]李东兵,李国平,滕国伟,等.一种新的自适应维纳滤波方法[J].上海大学学报:自然科学版,2012, 18(6):555-560.
[7]刘红亮,陈维义,许中胜.一种新的小波域维纳滤波图像去噪算法[J].海军工程大学学报,2015, 27(6):63-67.
[8]ZHANG X, FENG X. Multiple-step local Wiener filter with proper stopping in wavelet domain[J]. Journal of Visual Communication&Image Representation,2014,25(2):254-262.
[9]颜军,郝雪,王兆国.Wiener滤波器在小波变换域的应用[J].吉林大学学报,2006(s2):67-69.
[10] LAHMIRI S. An iterative denoising system based on Wiener filtering with application to biomedical images[J]. Optics&Laser Technology,2017, 90:128-132.
[11] IMMERKETR J. Fast noise variance estimation[J]. Computer Vision&Image Understanding, 1996, 64(2):300-302.