基于分层滤波技术的冲击噪声检测与去除算法
An Algorithm about Impulse Noise Detection and Removal Based on the Hierarchical Filter Technology

作者: 周颖玥 , 臧红彬 :西南科技大学信息工程学院,特殊环境机器人技术四川省重点实验室,四川 绵阳;

关键词: 图像去噪分层策略中值滤波噪声检测Image Denoising Hierarchical Stratagem Median Filter Noise Detection

摘要:
本文提出了一种基于分层滤波技术的图像随机值冲击噪声检测和去除算法。为了解决经典自适应中心加权中值滤波技术在较高噪声比率下漏检严重的问题,我们采取逐级设置由高到低的噪声值判决门限来层层筛选噪声,利用不同高度的门限甄别出不同信赖度的冲击噪点,同时更新已检测出的噪声点值。每一层均得到一幅中间去噪图像,并继续进行噪声点的判决,最终得到完整的噪声点空间以及终极去噪图像。通过对多幅不同噪声比率的噪声图进行图像恢复的仿真实验验证后,得出结论:本文所提算法无论在低噪声比率还是高噪声比率的情况下均能有效地检测出图像中的随机值冲击噪声,漏检和误检像素数相对较低,同时获得了较好的图像去噪效果,极大地扩展了原算法的应用范围。

Abstract: In this paper, an algorithm about random-valued impulse noise detection and removal based on the hierarchical filter technology is proposed. In order to solve the problem of severe miss detection of classical adaptive center-weighted median filter in the situation of high noise ratios, we set the noise judgment thresholds from high values to low values to select noisy pixels hierarchically. The impulse noise with different reliable degrees is selected through different thresholds and at the same time the detected pixel values are updated. The interim denoised image is generated in each denoising layer and the noise detection continues in the next denoising layer. Lastly, we get all the locations of noisy pixels and the final denoised image is obtained. Extensive experimental results for different noisy images with different noise ratios show that our proposed algorithm can obtain the good performance of random-valued impulse noise detection in the situations of not only low noise ratios but also high noise ratios. The miss and false noise detection ratios are both low relatively. At the same time, the good denoised images can be obtained. Our algorithm expands the application range of the classical adaptive center-weighted median filter.

文章引用: 周颖玥 , 臧红彬 (2015) 基于分层滤波技术的冲击噪声检测与去除算法。 图像与信号处理, 4, 27-35. doi: 10.12677/JISP.2015.43004

参考文献

[1] Bovik, A. (2000) Handbook of image and video processing. Academic Press, New York.

[2] Chen, T. and Wu, H.R. (2001) Space variant median filters for the restoration of impulse noise corrupted images. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, 48, 784-789.
http://dx.doi.org/10.1109/82.959870

[3] Garnett, R., Huegerich, T., Chui, C. and He, W. (2005) A universal noise removal algorithm with an impulse detector. IEEE Transactions on Image Processing, 14, 1747-1754.
http://dx.doi.org/10.1109/TIP.2005.857261

[4] Gonzalez, R.C. and Woods, R.E. (2002) Digital Image processing. 2nd Edition, Prentice Hall, New York.

[5] Lu, C. and Chou, T. (2012) Denoising of salt-and-pepper noise corrupted image using modified directional-weighted- median filter. Pattern Recognition Letters, 33, 1287-1295.
http://dx.doi.org/10.1016/j.patrec.2012.03.025

[6] Chen, T. and Ren Wu, H. (2001) Adaptive impulse detection using center-weighted median filters. IEEE Signal Processing Letters, 8, 1-3.
http://dx.doi.org/10.1109/97.889633

[7] Dong, Y. and Xu, S. (2007) A new directional weighted median filter for removal of random-valued impulse noise. IEEE Signal Processing Letters, 14, 193-196.
http://dx.doi.org/10.1109/LSP.2006.884014

[8] Wu, J. and Tang, C. (2014) Random-valued impulse noise re-moval using fuzzy weighted non-local means. Signal, Image and Video Processing, 8, 349-355.
http://dx.doi.org/10.1007/s11760-012-0297-1

[9] Xiong, B. and Yin, Z. (2012) A universal denoising framework with a new impulse detector and nonlocal means. IEEE Transactions on Image Processing, 21, 1663-1675.
http://dx.doi.org/10.1109/TIP.2011.2172804

[10] Yan, M. (2011) Restoration of images corrupted by impulse noise using blind inpainting and l0 norm. Ph.D. Thesis, University of California, Los Angeles.

[11] Xiao, Y., Zeng T., Yu, J. and Ng, M.K. (2011) Restoration of images corrupted by mixed Gaussian-impulse noise via l1-l0 minimization. Pattern Recognition, 44, 1708-1720.
http://dx.doi.org/10.1016/j.patcog.2011.02.002

[12] Wang, Z., Bovik, A.C., Sheikh, H.R. and Simonceli, E.P. (2004) Image quality assessment: From error visibility to structural similarity. IEEE Transactions on Image Processing, 13, 600-612.
http://dx.doi.org/10.1109/TIP.2003.819861

分享
Top