印刷电路板的图像分解去噪算法
|
摘要
为了改善印刷电路板(PCB)图像的视觉效果,提出基于图像分解的自适应加权L1范数和L2范数的PCB图像去噪算法。首先,将PCB噪声图像分解为结构和纹理两部分,其次设计一个自适应加权L1-L2范数正则化去噪模型。由于结构部分主要是分片平滑区域,体现PCB图像的整体框架,适合用L2范数各向同性去噪模型。纹理部分主要是高频信息,体现PCB图像的细节特征,适合用L1范数各向异性扩散正则化去噪模型。针对结构和纹理两个不同部分,设计自适应权函数,自动调整L1-L2范数正则化去噪模型中L1范数和L2范数的权值,然后,利用Bregman迭代算法得到最优的去噪效果。实验结果表明:与近年以来的相关经典去噪算法相比,利用新算法所得去噪图像的主观视觉效果更好,客观评价指标中的结构相似度可以提高27%以上,信噪比可以提高1 d B以上。
In order to improve the visual effect of printed circuit board( PCB) images,a PCB image denoising algorithm based on weighted L1 norm and L2 norm is proposed. Firstly,the PCB noise image is decomposed into two parts of structure and texture by using the non local mean image decomposition method. Secondly,a denoising model based on L1-L2 norm regularization is designed. The structure part embodies the whole frame of PCB image,which is mainly a piecewise smooth region. Therefore,the L2 norm isotropic diffusion regularization denoising model based on the high weight value is used to denoise it. The texture part embodies the detail features of PCB images,corresponds the high frequency information,so the L1 norm anisotropic diffusion regularization denoising model based on high weight is adopted to denoise it. Then,the Bregman iterative algorithm is applied to get the best denoising effect. The experimental results show that compared with the classical denoising algorithm in recent years,the subjective visual effect of the denoised image obtained by the new algorithm is better,and the objective evaluation indexes,such as the signal to noise ratio and the structure similarity,are also improved accordingly.
In order to improve the visual effect of printed circuit board( PCB) images,a PCB image denoising algorithm based on weighted L1 norm and L2 norm is proposed. Firstly,the PCB noise image is decomposed into two parts of structure and texture by using the non local mean image decomposition method. Secondly,a denoising model based on L1-L2 norm regularization is designed. The structure part embodies the whole frame of PCB image,which is mainly a piecewise smooth region. Therefore,the L2 norm isotropic diffusion regularization denoising model based on the high weight value is used to denoise it. The texture part embodies the detail features of PCB images,corresponds the high frequency information,so the L1 norm anisotropic diffusion regularization denoising model based on high weight is adopted to denoise it. Then,the Bregman iterative algorithm is applied to get the best denoising effect. The experimental results show that compared with the classical denoising algorithm in recent years,the subjective visual effect of the denoised image obtained by the new algorithm is better,and the objective evaluation indexes,such as the signal to noise ratio and the structure similarity,are also improved accordingly.
引文
1 Wu H,Zhang X M,Kuang Y C,et al. Solder joint inspection based on neural network combined with generic algorithm[J]. Optik,2013,124(20):4110-4116
2 乔闹生,张奋,黎小琴.印刷电路板光板缺陷图像预处理研究[J].激光与光电子学进展,2015(2):142-147Qiao Naosheng,Zhang Fen,Li Xiaoqin. Defect image preprocessing of printed circuit board[J]. Laser&Optoelectronics Progress,2015(2):142-147
3 张桂梅,孙晓旭,刘建新.基于自适应投影算法的分数阶全变分去噪模型.模式识别与人工智能[J],2016,29(11):1009-1018Zhang G M. Sun X X,Liu J X. Fractional total variation denoising model based on adaptive projection algorithm[J]. Pattern Recognition and Artificial Intelligence,2016,29(11):1009-1018
4 Jiang S,Hao X. Hybrid Fourier-wavelet image denoising[J]. Electronics Letters,2007,43(20):1081-1082
5 Othman H,Qian S E. Noise reduction of hyperspectral imagery using hybrid spatial-spectral derivative-domai wavelet shrinkage[J]. IEEE Trans on Geoscience and Remote Sensing,2006,44(2):397-408
6 蔡长青,张永山.改进阈值的窗口傅里叶变换滤波[J].激光与光电子学进展,2015(03):171-176Cai C Q,Zhang Y S. Windowed fourier transform filter method with improved threshold[J]. Laser&Optoelectronics Progress,2015(3):171-176
7 倪林,Miao Y.一种更适合图像处理的多尺度变换—Curvelet变换[J].计算机工程与应用,2004,28:21-26Ni L,Miao Y. Curvelet transform and its applications in image processing[J]. Computer Engineering and Applications,2004,28:21-26
8 周先春,吴婷,石兰芳,等.一种基于曲率变分正则化的小波变换图像去噪算法[J].电子学报,2018,46(3):621-628Zhou XC,Wu T,Shi LF,et al. A kind of wavelet transform image denoising method based on curvature variation regularization[J]. Acta Electronica Sinica,2018,46(3):621-628
9 Jean-Luc Starck,Emmanuel J. The curvelet transform for image denoising[J]. IEEE Transaction Image Processing,2002,11(6):670-684
10 Martin-Fernandez M. Sequential anisotropic wiener filtering applied to 3D MRI data[J]. Magnetic Resonance Imaging,2007,25(2):278-292
11 Jiang W L,Li G L,Luo W B. Application of improved median filtering algorithm to image de-noising[J]. Journal of Applied Optics,2011,998-999(4):838-841
12 Lecun Y,Bengio Y,Hinton G. Deep learning[J]. Nature,2015,521(7553):436-444
13 Zhang K,Zuo W,Chen Y,et al. Beyond a Gaussian denoiser:residual learning of deep CNN for image denoising[J]. IEEE Transaction on Image Processing,2017,26(7):3142-3155
14 Zhou Z F,Cao J Z,Wang H,et al. Image Denoising algorithm via doubly bilateral filtering[C]//International Conference on Information Engineering and Computer Science. New York:IEEE,2009:1-4
15 Buades A,Coll B,More L J M. A review of image denoising algorithm with a new one[J]. SIAM Journal on Multiscale Modeling and Simulation,2005,4(2):490-530
16 Perona P,Malik J. Scale-space and edge detection using anisotropic diffusion[J]. IEEE Transaction on Pattern Analysis and Machine Intelligence,1990,12(7):629-639
17 Prasath V B S,Vorotnikov D. On a system of adaptive coupled PEDs for image restoration[J]. Journal of Mathematical Imaging and Vision,2014,48(1):35-52
18 Tomasi C,Manduchi R. Bilateral filtering for gray and color images[C]//International Conference on Computer Vision. New York:IEEE,1998:839-846
19 卢碧波,王乐蓉.全变分引导的双边滤波图像去噪方法[J].光学技术,2018,44(2):194-200Lu Bibo,Wang Lerong. Total variation guided bilateral filtering for image denoising[J]. Optical Technique,2018,44(2):194-200
20 张闯,迟健男,张朝晖,等.基于边缘检测与双边滤波的彩色图像去噪[J].电子学报,2010,38(8):1776-1783Zhang C,Chi J N,Zhang Z H,et al. Removing noise of color images based on edge detection and bilateral filter[J]. Acta Electronica Sinica,2010,38(8):1776-1783
21 Buades A,Coll B,More L J M. A review of image denoising algorithm with a new one[J]. SIAM Journal on Multiscale Modeling and Simulation,2005,4(2):490-530
22 张俊峰,罗立民,舒华忠,等.基于先验信息的全变分图像复原算法[J].东南大学学报(自然科学版),2016,46(6):1132-1136Zhang J F,Luo L M,Shu H Z,et al. Total variation image restoration algorithm based on prior information[J]. Journal of Southeast University(Natural Science Edition),2016,46(6):1132-1136
23 Chen F,Zeng X X,Wang M Q. Image denoising via local and nonlocal circulant similarity[J]. Journal of Visual Communication and Image Representation,2015,30:117-124
24 王倩,彭海云,秦杰,等. NSCT域分类预处理的改进非局部均值去噪算法[J].计算机辅助设计与图形学学报,2018,30(3):436-446Wang Q,Peng HY,Qin J,et al. Improved NLM denoising algorithm based on classification preprocess in NSCT domain[J]. Journal of Computer Aided Design&Computer Graphics,2018,30(3):436-446
25 冯象初,刘鑫,杨春雨,等. L1范数约束的非局部均值正则图像去模糊模型[J].北京邮电大学学报,2018,41(1):81-87Feng X C,Liu X,Yang C Y,et al. L1-Nonlocal means regularization model for image deblurring problem[J]. Journal of Beijing University of Posts and Telecommunications,2018,41(1):81-87
26 Rudin L,Osher S,Fatemi E. Nonlinear total variation based noise removal algorithms[J]. Physica D,1992,60(1-4):259-268
27 张红英,彭启琮.全变分自适应去噪模型[J].光电工程,2006,33(3):50-53Zhang HY,Peng QC. Adaptive image denoising model based on total variation[J]. Opto-Electronic Engineering,2006,33(3):50-53
28 何坤,郑秀清,琚生根,等.改进全变分的图像去噪[J].电子科技大学学报,2016,45(3):463-468He K,Zheng XQ,Ju SG,et al. Image denoising on improved total variation[J]. Journal of University of Electronic Science and Technology of China,2016,45(3):463-468
29 Della P S,Della P V,Lafferty J. Duality and auxiliary functions for Bregman distances(CMU-CS-01-109)[R]. School of Computer Science Carnegie Mellon University,2002(2):1-16
30 Ng M K,Yuan X,Zhang W. Coupled variational image decomposition and restoration model for blurred cartoonplus-texture images with missing pixels[J]. IEEE Transactions on Image Processing,2013,22(6):2233-2246
31 Stanley O,Martin B,Donald G,et al. An iterative regularization method for total variation-based image restoration[J]. Multiscale Modeling and Simulation,2005,4(2):460-489
32 Wang Z,Bovik AC,et al. Image quality assessment:from error visibility to structural similarity[J]. IEEE Transactions on Image Processing,2004,13(4):600-612
2 乔闹生,张奋,黎小琴.印刷电路板光板缺陷图像预处理研究[J].激光与光电子学进展,2015(2):142-147Qiao Naosheng,Zhang Fen,Li Xiaoqin. Defect image preprocessing of printed circuit board[J]. Laser&Optoelectronics Progress,2015(2):142-147
3 张桂梅,孙晓旭,刘建新.基于自适应投影算法的分数阶全变分去噪模型.模式识别与人工智能[J],2016,29(11):1009-1018Zhang G M. Sun X X,Liu J X. Fractional total variation denoising model based on adaptive projection algorithm[J]. Pattern Recognition and Artificial Intelligence,2016,29(11):1009-1018
4 Jiang S,Hao X. Hybrid Fourier-wavelet image denoising[J]. Electronics Letters,2007,43(20):1081-1082
5 Othman H,Qian S E. Noise reduction of hyperspectral imagery using hybrid spatial-spectral derivative-domai wavelet shrinkage[J]. IEEE Trans on Geoscience and Remote Sensing,2006,44(2):397-408
6 蔡长青,张永山.改进阈值的窗口傅里叶变换滤波[J].激光与光电子学进展,2015(03):171-176Cai C Q,Zhang Y S. Windowed fourier transform filter method with improved threshold[J]. Laser&Optoelectronics Progress,2015(3):171-176
7 倪林,Miao Y.一种更适合图像处理的多尺度变换—Curvelet变换[J].计算机工程与应用,2004,28:21-26Ni L,Miao Y. Curvelet transform and its applications in image processing[J]. Computer Engineering and Applications,2004,28:21-26
8 周先春,吴婷,石兰芳,等.一种基于曲率变分正则化的小波变换图像去噪算法[J].电子学报,2018,46(3):621-628Zhou XC,Wu T,Shi LF,et al. A kind of wavelet transform image denoising method based on curvature variation regularization[J]. Acta Electronica Sinica,2018,46(3):621-628
9 Jean-Luc Starck,Emmanuel J. The curvelet transform for image denoising[J]. IEEE Transaction Image Processing,2002,11(6):670-684
10 Martin-Fernandez M. Sequential anisotropic wiener filtering applied to 3D MRI data[J]. Magnetic Resonance Imaging,2007,25(2):278-292
11 Jiang W L,Li G L,Luo W B. Application of improved median filtering algorithm to image de-noising[J]. Journal of Applied Optics,2011,998-999(4):838-841
12 Lecun Y,Bengio Y,Hinton G. Deep learning[J]. Nature,2015,521(7553):436-444
13 Zhang K,Zuo W,Chen Y,et al. Beyond a Gaussian denoiser:residual learning of deep CNN for image denoising[J]. IEEE Transaction on Image Processing,2017,26(7):3142-3155
14 Zhou Z F,Cao J Z,Wang H,et al. Image Denoising algorithm via doubly bilateral filtering[C]//International Conference on Information Engineering and Computer Science. New York:IEEE,2009:1-4
15 Buades A,Coll B,More L J M. A review of image denoising algorithm with a new one[J]. SIAM Journal on Multiscale Modeling and Simulation,2005,4(2):490-530
16 Perona P,Malik J. Scale-space and edge detection using anisotropic diffusion[J]. IEEE Transaction on Pattern Analysis and Machine Intelligence,1990,12(7):629-639
17 Prasath V B S,Vorotnikov D. On a system of adaptive coupled PEDs for image restoration[J]. Journal of Mathematical Imaging and Vision,2014,48(1):35-52
18 Tomasi C,Manduchi R. Bilateral filtering for gray and color images[C]//International Conference on Computer Vision. New York:IEEE,1998:839-846
19 卢碧波,王乐蓉.全变分引导的双边滤波图像去噪方法[J].光学技术,2018,44(2):194-200Lu Bibo,Wang Lerong. Total variation guided bilateral filtering for image denoising[J]. Optical Technique,2018,44(2):194-200
20 张闯,迟健男,张朝晖,等.基于边缘检测与双边滤波的彩色图像去噪[J].电子学报,2010,38(8):1776-1783Zhang C,Chi J N,Zhang Z H,et al. Removing noise of color images based on edge detection and bilateral filter[J]. Acta Electronica Sinica,2010,38(8):1776-1783
21 Buades A,Coll B,More L J M. A review of image denoising algorithm with a new one[J]. SIAM Journal on Multiscale Modeling and Simulation,2005,4(2):490-530
22 张俊峰,罗立民,舒华忠,等.基于先验信息的全变分图像复原算法[J].东南大学学报(自然科学版),2016,46(6):1132-1136Zhang J F,Luo L M,Shu H Z,et al. Total variation image restoration algorithm based on prior information[J]. Journal of Southeast University(Natural Science Edition),2016,46(6):1132-1136
23 Chen F,Zeng X X,Wang M Q. Image denoising via local and nonlocal circulant similarity[J]. Journal of Visual Communication and Image Representation,2015,30:117-124
24 王倩,彭海云,秦杰,等. NSCT域分类预处理的改进非局部均值去噪算法[J].计算机辅助设计与图形学学报,2018,30(3):436-446Wang Q,Peng HY,Qin J,et al. Improved NLM denoising algorithm based on classification preprocess in NSCT domain[J]. Journal of Computer Aided Design&Computer Graphics,2018,30(3):436-446
25 冯象初,刘鑫,杨春雨,等. L1范数约束的非局部均值正则图像去模糊模型[J].北京邮电大学学报,2018,41(1):81-87Feng X C,Liu X,Yang C Y,et al. L1-Nonlocal means regularization model for image deblurring problem[J]. Journal of Beijing University of Posts and Telecommunications,2018,41(1):81-87
26 Rudin L,Osher S,Fatemi E. Nonlinear total variation based noise removal algorithms[J]. Physica D,1992,60(1-4):259-268
27 张红英,彭启琮.全变分自适应去噪模型[J].光电工程,2006,33(3):50-53Zhang HY,Peng QC. Adaptive image denoising model based on total variation[J]. Opto-Electronic Engineering,2006,33(3):50-53
28 何坤,郑秀清,琚生根,等.改进全变分的图像去噪[J].电子科技大学学报,2016,45(3):463-468He K,Zheng XQ,Ju SG,et al. Image denoising on improved total variation[J]. Journal of University of Electronic Science and Technology of China,2016,45(3):463-468
29 Della P S,Della P V,Lafferty J. Duality and auxiliary functions for Bregman distances(CMU-CS-01-109)[R]. School of Computer Science Carnegie Mellon University,2002(2):1-16
30 Ng M K,Yuan X,Zhang W. Coupled variational image decomposition and restoration model for blurred cartoonplus-texture images with missing pixels[J]. IEEE Transactions on Image Processing,2013,22(6):2233-2246
31 Stanley O,Martin B,Donald G,et al. An iterative regularization method for total variation-based image restoration[J]. Multiscale Modeling and Simulation,2005,4(2):460-489
32 Wang Z,Bovik AC,et al. Image quality assessment:from error visibility to structural similarity[J]. IEEE Transactions on Image Processing,2004,13(4):600-612