数字图像高密度脉冲噪声的中值滤波算法研究
|
摘要
图像在采集和传输过程中，往往会引入不同程度的噪声，这就为后面的边缘检测、图像分割和形状识别等带来很大的难度，所以图像去噪是图像处理中非常重要的一环。影响图像质量的噪声中最常见的是高斯噪声和脉冲噪声，本文主要针对脉冲噪声提出新的滤波算法。中值滤波及其各种改进方法是常用的去除脉冲噪声的滤波方法，它对去除脉冲噪声具有较好的性能，去噪后在人眼敏感的平滑区域不存在或有较少的噪声，但是现存的中值滤波算法在图像的结构细节保护上较差，尤其是在高密度脉冲噪声污染下，复原引起的失真较严重。
本文针对高密度脉冲噪声提出两种改进型的中值滤波算法。一种滤波方法命名为HsADMF是针对高密度椒盐噪声污染情况下的改进中值滤波法。HsADMF是基于序列图像的椒盐噪声去除的算法，利用椒盐噪声的正负脉冲特性，提出点对点的检测算法，把像素点分为信息点和可疑噪声点，仅对噪声点进行滤波处理，充分利用每幅图像的有用信息来恢复受污染的图像，取得了良好的恢复效果。分灰度图像和彩色图像分析该滤波方法的特性及其滤波效果图。第二种滤波方法命名为sADMF-VMF是根据HsADMF滤波方法的思想，提出针对高密度随机值脉冲噪声的滤波处理方法。该旅波方法首先对图像效果影响较大的极限值噪声用HsADMF方法处理，在此基础上再对随机脉冲噪声进行矢量中值滤波处理。两种滤波方法得到的峰值信噪比(PsNR)与其他经典的滤波方法相比最大，平均绝对误差(MAE)最小，即其去噪效果最好，细节处理效果最佳。
Digital images are often corrupted by the impulse noise in the image acquisitions and the image transmission progress because of the sensor noise and channel noise. This may result in difficulties in the subsequent image processing, such as edge detection, image segmentation and object recognition, therefore, detection and removal of the noise are inevitable in the image processing. The Gauss noise and impulse noise are the two most commonly used noise models. This thesis is mainly concerned in the development on the filtering algorithms for the impulse noise.
Median filter and its improved algorithms are normally considered to have good performances on removing impulse noise. After the corrupted images are filtered by the median filter, the smooth areas in the filtered images will contain no or little noise. However, the existing median filtering algorithms normally have poor performances on protecting the image details. Especially these algorithms may cause serious distortion in the images which are highly corrupted by the impulse noise.
In this thesis two improved median filtering algorithms are presented to remove high-ratio impulse noise. One filter (named HSADMF) is focused on filtering color images highly corrupted by the salt-and-pepper impulse noise. This algorithm firstly detects and processes the salt-and-pepper impulse noise using point to point method according to the characteristics of this noise. Then HSADMF takes full use of the oriental image information to filter the corrupted one. The performances of HSADMF on filtering gray and color images are compared with those of some classic filters. The comparison shows that HSADMF has better performances on detail protection and higher restoration ratio. The other filter (named SADMF-VMF) is focused on filtering color images highly corrupted by the random impulse noise. SADMF-VMF can be divided into two steps to remove random impulse noise, firstly it used HSADMF to remove salt and pepper noise which put a greater impact on the quality of the image, secondly it suppress other corrupted pixels by VMF. The performances of SADMF-VMF on filtering color images are compared with those of some classic filters. The comparison shows that SADMF-VMF has better performances on detail protection and higher restoration ratio.
本文针对高密度脉冲噪声提出两种改进型的中值滤波算法。一种滤波方法命名为HsADMF是针对高密度椒盐噪声污染情况下的改进中值滤波法。HsADMF是基于序列图像的椒盐噪声去除的算法，利用椒盐噪声的正负脉冲特性，提出点对点的检测算法，把像素点分为信息点和可疑噪声点，仅对噪声点进行滤波处理，充分利用每幅图像的有用信息来恢复受污染的图像，取得了良好的恢复效果。分灰度图像和彩色图像分析该滤波方法的特性及其滤波效果图。第二种滤波方法命名为sADMF-VMF是根据HsADMF滤波方法的思想，提出针对高密度随机值脉冲噪声的滤波处理方法。该旅波方法首先对图像效果影响较大的极限值噪声用HsADMF方法处理，在此基础上再对随机脉冲噪声进行矢量中值滤波处理。两种滤波方法得到的峰值信噪比(PsNR)与其他经典的滤波方法相比最大，平均绝对误差(MAE)最小，即其去噪效果最好，细节处理效果最佳。
Digital images are often corrupted by the impulse noise in the image acquisitions and the image transmission progress because of the sensor noise and channel noise. This may result in difficulties in the subsequent image processing, such as edge detection, image segmentation and object recognition, therefore, detection and removal of the noise are inevitable in the image processing. The Gauss noise and impulse noise are the two most commonly used noise models. This thesis is mainly concerned in the development on the filtering algorithms for the impulse noise.
Median filter and its improved algorithms are normally considered to have good performances on removing impulse noise. After the corrupted images are filtered by the median filter, the smooth areas in the filtered images will contain no or little noise. However, the existing median filtering algorithms normally have poor performances on protecting the image details. Especially these algorithms may cause serious distortion in the images which are highly corrupted by the impulse noise.
In this thesis two improved median filtering algorithms are presented to remove high-ratio impulse noise. One filter (named HSADMF) is focused on filtering color images highly corrupted by the salt-and-pepper impulse noise. This algorithm firstly detects and processes the salt-and-pepper impulse noise using point to point method according to the characteristics of this noise. Then HSADMF takes full use of the oriental image information to filter the corrupted one. The performances of HSADMF on filtering gray and color images are compared with those of some classic filters. The comparison shows that HSADMF has better performances on detail protection and higher restoration ratio. The other filter (named SADMF-VMF) is focused on filtering color images highly corrupted by the random impulse noise. SADMF-VMF can be divided into two steps to remove random impulse noise, firstly it used HSADMF to remove salt and pepper noise which put a greater impact on the quality of the image, secondly it suppress other corrupted pixels by VMF. The performances of SADMF-VMF on filtering color images are compared with those of some classic filters. The comparison shows that SADMF-VMF has better performances on detail protection and higher restoration ratio.
引文
[1]．孙即祥编著。图像处理。科学出版社，2004：2-6
[2]．I Pitas，A N Venetsanopoulos．Nonlinear digital filters：principles and applications．Boston，MA：Kluwer,1990．
[3]．葛亭，几种数字图像滤波算法，南京信息工程大学学报，2006，07：2-3。
[4]．章毓晋，图像工程(上册)，图像处理与分析，清华大学出版社，1999，2：135-158
[5]．DigitallmageProeessing(SeeondEdition)，RafaelC．Gonsalesz，Rieharde．Woods，Publishing House of Eleetronies lndustry．Beijing，：13-14
[6]．余成波．数字图像处理及MATLAB实现．重庆大学出版社，2002
[7]．张兆礼，赵春晖，梅晓丹。现代图像处理技术及Matlab实现。人民邮电出版社，2001
[8]．李永强，沈庆国，朱江，汪莉，数字视频质量评价方法综述，节目与广播，2006 Vol．6，74-82
[9]．Arce C．R，Gallagher N．C．State Description for the root signal set of median filters．IEEE Trans Acoust，Speech，Signal Processing．1982，30：894-902．
[10]．Wendt P．D，Coyle E．J，Gallagher N．C．Stack Filters．IEEE Trans Acoust，Speech，Signal Processing．1986，34：898-911．
[11]．Lee Y H．．Kassam S．A．Generalized Median Filtering Related Nonlinear filtering Techniques．IEEE Trans Acoust，Speech，Signal Processing．1985，33(3)：672-683．
[12]．Bovik A．C，Huang T．S，Munson D．C．A Generalization of Median Filters Using Liner Combination of Order Statistics．IEEE Trans Acoust，Speech，Signal Processing，1983，31(6)：1342-1350．
[13]．R Bemstein．Adaptive Nonlinear Filters for Simultaneous Removal of Different Kinds of Noise in Image．IEEE Trans Circuits and Systems，1987，34(11)：1275-1291．
[14]．L．Arslan，A．McCree，V viswanathan．New Methods for Adaptive Noise Suppression．Proceedings of IEEE International Conference on Acoustics，Speech，and Signal Processing．Los Alamitos，California：IEEE Computer Society Press，1995，1：812-815．
[15]．D Brownrigg．The weighted median filter[j]．Commu．Assoc．Computer,1984，27(8)：807-818
[16]．S．J Ko，S J Lee．Center weighted median filters and their applications to image enchancement．IEEE Trans．Circuits Syst．1991，15：984-993．
[17]．J．Zheng，K．P Valavanis，J．M．Gauch-Noise removal form color images．J．Intell．Robot．Syst．7(3)(June 1993)：257-285．
[18]．T Sun，Y Neuvo．Detail-preserving median．Based filters in image processing[J]．Pattern Recognition Letters，1994，15(4)：341-347．
[19]．董继扬，张军英。严重椒盐噪声污染图像的非线性滤波算法。光电子激光，2003，14(12)：1336-1339。
[20]．G R Arce，R E Roster.Multilevel median filter：properties and efficacy．Proc．ICASSP,New York，1998．
[21]．Lianghai Jin，Dehua Li．A switching vector median filter based on the CIELAB color space for color image restoration．Signal Processing 87(2007)：1345-1354．
[22]．R．Lukac，B．Smolka，K．Martin，K．N．Platatniotis，A．N．Venetsanopoulos．Vector filtering for color imaging．IEEE signal Process．Mag，22(1)(January 2005)：74-86．
[23]．J．Astola，R Haavisto，Y Neuvo．Vector median filters．Proc．IEEE78(4)(April 1990)：678-689．
[24]．P．E．Trahanias，A．N．Venetsanopoulos，Vector directional filters：a new class of multichannel image processing filters．IEEE Trans．Image Process．2(4)(October 1993)：528-534．
[25]．P．E．Trahanias，D．G Karakos，A．N．Venetsanopoulos．Directional processing of color images：theory and ecperimental results．IEEE Trans．Image Process．5(6)(June 1996)：868-880．
[26]．Karakos D．G，Tranias P．E．Combining vector median and vector directional filters：the directional-distance filters．Proc ofIEEE Int．Confon Image Processing 95．1995：171-174．
[27]．蒋刚毅，郁梅，郁伯康．自适应猜测图像滤波新方法．科学通报，1999，44(18)：1937-1942．
[28]．蒋刚毅，郁梅，郁伯康，等．彩色形态滤波起分析．电子学报，1999，27(11)：57-60．
[29]．郭晓新，卢奕南，许志闻，王云霄，庞云阶．自适应定向加权中值滤波吉林大学学报，2005，43(4)，：493-496。
[30]．E Abreu，M Lightson，S．K．Mitra，et al．A new efficient approach for the removal of impulse noise form highly corrupted images．IEEE Trans．Image Processing，digital signal processing，1999，46(1)：78-80．
[31]．W Han，J lin．Minimum-maximum exclusive mean(MMEM)filter to remove impulse noise form highly corrupted images．Electronics Lett．，1997，33(2)：124-125．
[32]．Karakos D．D，Trahanias P E，Generalized multichannel image-filtering structures，IEEE Transactions on Image Processing，1997，6(7)：1038-1045．
[33]．Wang Zhou，Zhang David，Progressive switching median filter for the removal of impulse noise form highly corrupted images，IEEE Trans．On Circuits and Systems，Analog and Digital Signal Processing，1999，10(2)：655-659
[34]．穆静。对图像消噪的中值滤波器的改进，现代电子技术。2004(15)：21-25。
[35]．王建华，王春平，贾洪涛。自适应中值滤波其在图像降噪技术中的应用。测控技术，2004，23(5)：54-56。
[36]．董继扬，张军英，一种简单的椒盐噪声滤波算法，计算机工程与应用，2003(20)：63-66。
[37]．张旭明，徐滨士，董世运，甘小明，自适应中值一加权均值混合滤波器。光学技术，2004，30(6)：652-659。
[38]．Xiaowei Nan，Junsheng Li，et al，A selective and adaptive image filtering approach based on impulse noise detection．Proceeding of the 5th world Congress on Intelligent Control and Automation．2004．6：15-19．
[39]．T．sun，Y．Neuvo，Detail-preserving median based filters in image processing，Pattern Recognit Letters，1994，15(4)：341-347．
[40]．T．chen，K．-K．Ma，L．-H．chen，Tri-state median filter for image denoising．IEEE Trans Image Process，1999，8(2)：1834-1838．
[41]．T．Lin，R Yu，Thresholding Noise-flee Ordered Mean filtering based on Dempster-Shafer theoryfor image restoration．IEEE Trans，CAS-II，2006，53(5)：1057-1064．
[42]．w．Luo，An efficient detail-preserving approach for removing impulse noise in images．IEEE SPL，2006，13(7)：413-416．
[43]．R．Chan，C．HLl，M．Nikolova，Salt and Pepper noise removal by median-type noise detectors and detml-preserving regularizmion，IEEE Trans on Ip,2006，14(10)：1479-1485．
[44]．Tao Chen and Hongren Wu，Adaptive Impulse Detection Using Center-Weighted Median Filters．IEEE signal processing letters，2001，8(1)：1-3．
[45]．R．Chan，C，Hu，M．Nikolova，An iterative procedure for removing random-valued impulse noise．IEEE，SP LeUers，2004，11(12)：921-924．
[46]．J．Chang，J．Chen，Classifier-augmented median filters for image restoration，IEEE，Trans on Instrumentation and measurement，2004，53(2)：351-356．
[47]．I．Aizenberg，C．Butakoif,Effective impulse detector based on rank-order criteria，IEEE，SP Letters，2004，11(3)：363-366．
[48]．I．Andreadis，G Louverdis，Real-time adaptive image impulse noise suppression．IEEE Trans on instrumentation and measurement，2004，53(3)：798-806．
[49]．R S．Windyga．Fast Impursive noise Removal，IEEE Transactions on Image processing，2001，10(1)：173-179．
[50]．Z．Wang，D．Zhang Progressive Switching Median Filter for the Removal of Impulse Noise from Highily corrupted Images，IEEE Transaction on Circuits and Systems，1999，46(1)：78-80．
[2]．I Pitas，A N Venetsanopoulos．Nonlinear digital filters：principles and applications．Boston，MA：Kluwer,1990．
[3]．葛亭，几种数字图像滤波算法，南京信息工程大学学报，2006，07：2-3。
[4]．章毓晋，图像工程(上册)，图像处理与分析，清华大学出版社，1999，2：135-158
[5]．DigitallmageProeessing(SeeondEdition)，RafaelC．Gonsalesz，Rieharde．Woods，Publishing House of Eleetronies lndustry．Beijing，：13-14
[6]．余成波．数字图像处理及MATLAB实现．重庆大学出版社，2002
[7]．张兆礼，赵春晖，梅晓丹。现代图像处理技术及Matlab实现。人民邮电出版社，2001
[8]．李永强，沈庆国，朱江，汪莉，数字视频质量评价方法综述，节目与广播，2006 Vol．6，74-82
[9]．Arce C．R，Gallagher N．C．State Description for the root signal set of median filters．IEEE Trans Acoust，Speech，Signal Processing．1982，30：894-902．
[10]．Wendt P．D，Coyle E．J，Gallagher N．C．Stack Filters．IEEE Trans Acoust，Speech，Signal Processing．1986，34：898-911．
[11]．Lee Y H．．Kassam S．A．Generalized Median Filtering Related Nonlinear filtering Techniques．IEEE Trans Acoust，Speech，Signal Processing．1985，33(3)：672-683．
[12]．Bovik A．C，Huang T．S，Munson D．C．A Generalization of Median Filters Using Liner Combination of Order Statistics．IEEE Trans Acoust，Speech，Signal Processing，1983，31(6)：1342-1350．
[13]．R Bemstein．Adaptive Nonlinear Filters for Simultaneous Removal of Different Kinds of Noise in Image．IEEE Trans Circuits and Systems，1987，34(11)：1275-1291．
[14]．L．Arslan，A．McCree，V viswanathan．New Methods for Adaptive Noise Suppression．Proceedings of IEEE International Conference on Acoustics，Speech，and Signal Processing．Los Alamitos，California：IEEE Computer Society Press，1995，1：812-815．
[15]．D Brownrigg．The weighted median filter[j]．Commu．Assoc．Computer,1984，27(8)：807-818
[16]．S．J Ko，S J Lee．Center weighted median filters and their applications to image enchancement．IEEE Trans．Circuits Syst．1991，15：984-993．
[17]．J．Zheng，K．P Valavanis，J．M．Gauch-Noise removal form color images．J．Intell．Robot．Syst．7(3)(June 1993)：257-285．
[18]．T Sun，Y Neuvo．Detail-preserving median．Based filters in image processing[J]．Pattern Recognition Letters，1994，15(4)：341-347．
[19]．董继扬，张军英。严重椒盐噪声污染图像的非线性滤波算法。光电子激光，2003，14(12)：1336-1339。
[20]．G R Arce，R E Roster.Multilevel median filter：properties and efficacy．Proc．ICASSP,New York，1998．
[21]．Lianghai Jin，Dehua Li．A switching vector median filter based on the CIELAB color space for color image restoration．Signal Processing 87(2007)：1345-1354．
[22]．R．Lukac，B．Smolka，K．Martin，K．N．Platatniotis，A．N．Venetsanopoulos．Vector filtering for color imaging．IEEE signal Process．Mag，22(1)(January 2005)：74-86．
[23]．J．Astola，R Haavisto，Y Neuvo．Vector median filters．Proc．IEEE78(4)(April 1990)：678-689．
[24]．P．E．Trahanias，A．N．Venetsanopoulos，Vector directional filters：a new class of multichannel image processing filters．IEEE Trans．Image Process．2(4)(October 1993)：528-534．
[25]．P．E．Trahanias，D．G Karakos，A．N．Venetsanopoulos．Directional processing of color images：theory and ecperimental results．IEEE Trans．Image Process．5(6)(June 1996)：868-880．
[26]．Karakos D．G，Tranias P．E．Combining vector median and vector directional filters：the directional-distance filters．Proc ofIEEE Int．Confon Image Processing 95．1995：171-174．
[27]．蒋刚毅，郁梅，郁伯康．自适应猜测图像滤波新方法．科学通报，1999，44(18)：1937-1942．
[28]．蒋刚毅，郁梅，郁伯康，等．彩色形态滤波起分析．电子学报，1999，27(11)：57-60．
[29]．郭晓新，卢奕南，许志闻，王云霄，庞云阶．自适应定向加权中值滤波吉林大学学报，2005，43(4)，：493-496。
[30]．E Abreu，M Lightson，S．K．Mitra，et al．A new efficient approach for the removal of impulse noise form highly corrupted images．IEEE Trans．Image Processing，digital signal processing，1999，46(1)：78-80．
[31]．W Han，J lin．Minimum-maximum exclusive mean(MMEM)filter to remove impulse noise form highly corrupted images．Electronics Lett．，1997，33(2)：124-125．
[32]．Karakos D．D，Trahanias P E，Generalized multichannel image-filtering structures，IEEE Transactions on Image Processing，1997，6(7)：1038-1045．
[33]．Wang Zhou，Zhang David，Progressive switching median filter for the removal of impulse noise form highly corrupted images，IEEE Trans．On Circuits and Systems，Analog and Digital Signal Processing，1999，10(2)：655-659
[34]．穆静。对图像消噪的中值滤波器的改进，现代电子技术。2004(15)：21-25。
[35]．王建华，王春平，贾洪涛。自适应中值滤波其在图像降噪技术中的应用。测控技术，2004，23(5)：54-56。
[36]．董继扬，张军英，一种简单的椒盐噪声滤波算法，计算机工程与应用，2003(20)：63-66。
[37]．张旭明，徐滨士，董世运，甘小明，自适应中值一加权均值混合滤波器。光学技术，2004，30(6)：652-659。
[38]．Xiaowei Nan，Junsheng Li，et al，A selective and adaptive image filtering approach based on impulse noise detection．Proceeding of the 5th world Congress on Intelligent Control and Automation．2004．6：15-19．
[39]．T．sun，Y．Neuvo，Detail-preserving median based filters in image processing，Pattern Recognit Letters，1994，15(4)：341-347．
[40]．T．chen，K．-K．Ma，L．-H．chen，Tri-state median filter for image denoising．IEEE Trans Image Process，1999，8(2)：1834-1838．
[41]．T．Lin，R Yu，Thresholding Noise-flee Ordered Mean filtering based on Dempster-Shafer theoryfor image restoration．IEEE Trans，CAS-II，2006，53(5)：1057-1064．
[42]．w．Luo，An efficient detail-preserving approach for removing impulse noise in images．IEEE SPL，2006，13(7)：413-416．
[43]．R．Chan，C．HLl，M．Nikolova，Salt and Pepper noise removal by median-type noise detectors and detml-preserving regularizmion，IEEE Trans on Ip,2006，14(10)：1479-1485．
[44]．Tao Chen and Hongren Wu，Adaptive Impulse Detection Using Center-Weighted Median Filters．IEEE signal processing letters，2001，8(1)：1-3．
[45]．R．Chan，C，Hu，M．Nikolova，An iterative procedure for removing random-valued impulse noise．IEEE，SP LeUers，2004，11(12)：921-924．
[46]．J．Chang，J．Chen，Classifier-augmented median filters for image restoration，IEEE，Trans on Instrumentation and measurement，2004，53(2)：351-356．
[47]．I．Aizenberg，C．Butakoif,Effective impulse detector based on rank-order criteria，IEEE，SP Letters，2004，11(3)：363-366．
[48]．I．Andreadis，G Louverdis，Real-time adaptive image impulse noise suppression．IEEE Trans on instrumentation and measurement，2004，53(3)：798-806．
[49]．R S．Windyga．Fast Impursive noise Removal，IEEE Transactions on Image processing，2001，10(1)：173-179．
[50]．Z．Wang，D．Zhang Progressive Switching Median Filter for the Removal of Impulse Noise from Highily corrupted Images，IEEE Transaction on Circuits and Systems，1999，46(1)：78-80．