基于矢量中值的彩色图像中脉冲噪声去除算法的研究
|
摘要
随着多媒体的应用日益广泛。图像,已日益深入到人们的生活中。图像在采集、传输、解码过程中都有可能产生噪声。噪声使图像出现颜色的失真,清晰度下降,影响图像质量。由于产生的原因不同,噪声有多种类型,其中脉冲噪声最为常见,对图像质量有很大的影响。
本文针对彩色图像脉冲噪声的去除算法进行了研究。根据彩色图像的特性,选择了结合了矢量中值滤波器(VMF)和矢量方向滤波器(VDF)的优点的方向-距离滤波器(DDF)作为基础。并提出了两种改进的噪声去除算法。第一种算法首先进行了区域检测,然后使用第一种噪声检测算子,最后使用高斯加权窗的DDF算法,处理被噪声污染的点;第二种算法是在第一种算法的基础上改进的,它的特点是运算量比第一种改进算法小,而且使用第二种检测算子,将疑为噪声点分为孤立噪声点和细节区,并利用VMF和DDF的优点,对上述区域分别使用VMF和高斯加权窗的DDF算法。
仿真结果表明,使用了第一种改进的噪声去除算法的处理之后,图像的噪声去除效果较好,而且细节和色度保持很好。而且,使用了高斯加权窗的DDF方法比经典的VMF和VDF有更好的细节保持效果和较好的色度保持效果;而第二种改进算法,相比第一种改进算法是一种快速检测算法,目的是为了在图像中区分孤立噪声点和细节区,并利用VMF和DDF各自的优点。实验结果证明,第二种改进算法也具有较好的噪声去除效果。
The color image processing system has been employed in many fields, from image-capture, to feature-abstraction, and the most common processing technique is filtering. And for the most cases, the noises are additive. Filtering is of the most importance. While during the filtering techniques, the most common one is the linear filter, and it is well-known for the simple mathematic expression, the uniform theory basics and simple implementation. However, most of them are on the base of Gauss statistics model.
But, many questions could not been solved by the linear filtering technology. For example, the linear technologies have been limited by the application of image processing. Traditional linear technologies could not become the model of image-forming or image-transportation, and it has not considered the nonlinear model of human vision system. The images compose of smooth fields and the detailed fields, especially, the details mean the better to vision system. The filters which have the characteristic of detail-preservation have gained more and more applications. But most of the linear processing systems will make the details and the borderlines vague.
The noises will degrade the comprehension of human to images, so it becomes the limited factor. Similarly, if you put the images to the quantitative analysis, the noises will restrict the performance of the systems. As a result, noises compression has been the key part of the image processing system, which restores the original images from the images polluted by the impulsive noises.
There are noises in the process of images forming, transportation, and the decoding. The noises make the images color distortion, definition degradation. Due to the different reasons of generation, there are several types of noises. And the most generic one is the impulsive noise, and its impact to images is not negligent. Transportation errors, random movements of camera, instability of electrons, electromagnetic disturbance, errors of sensors, it is common in the transportation in the airs, including the light, the mechanical movements, the motor engines, or high-voltage line errors, and all kinds of electronic switches. And these become the interferences to the broadcast and the communication signals, while the corresponding images will be noisy. But the huge errors will generate a kind of impulsive noises, which appears as black or white pot, so it is called Peppers and Salt noises.
The median filter is well-known for its effects of noise suppression to impulsive noise, but stretches the median filter to the color images simply, that is to imply the median to the RGB channel separately, will not generate good effects, since there are relations between the three channels. So the color distortion will appear. Due to these reasons, the vector median filters appear.
Because the shortcomings of general vector algorithms are: they consider the threads and details as noises, then to process as usual. The image quality will decline, too. Therefore, noise-detection algorithm is to identify the noises, fine details, at the same time, separate the marginal fields and the details from the points which were noise polluted. That is the aim of noise-detection. But the most important thing is to distinguish between lines or the details and the noisy points, so, we introduced the edge-detection algorithms, thereby, to suppress the noises. This paper summarizes the main representative algorithms of several edge- detection algorithms and noise-detection algorithms, firstly, introduced some common edge-detection algorithm, and then introduced several representative algorithms, the first is CANNY edge-detection operator, which is a kind of gradient algorithm operator, the second one is the CIELAB color space noise-detection algorithm.
To the gray images, the edges could be seen as the points whose gray scale jump, or they can be described as the gradual change, or the direction of the maximum gradient, in the gray-scale images, gradient operator can be used to detect image edges, but it is effective to the step edge only. When the gray width increases, the second derivative is a good method, instead, for example, the Laplace operator. And the edges could be described as follows: the zero crossing point of the second derivative. However, in the color image processing, color images are of the multi-channel, so vector processing method is more applicable.
Therefore, in this paper, two types improved noise detection operator have been proposed. The first type is based on the vector computing method, and the second kind is a rapid detection, after the detection, the images were dealt with different methods. Firstly, VMF and DDF filter have been chosen as a basis, but in order to choose a common filter window, so the simulation analysis has been implemented. In this paper, improved color image filtering algorithms have been divided into two types: the first type algorithm, the image of the first divided into different districts, then imply the first noises detection operator to detect the noises points, finally, to used the improved filtering window to the DDF, or VMF; and for the second type algorithm, it is the fast detection algorithm, and it deal with the noisy point and the details with different methods. In the first type algorithm, it is on the basis of the first improved noise-detection algorithms, because it has compared the pros and cons of VMF and DDF, therefore, in the second type of algorithm, firstly, to use the rapid detection, which detected the noisy points, at the same time to avoid unnecessary treatment, such as the threads, then the noisy points could be detected and divided into the isolated noisy points and the details fields, and finally treat these noisy points with different algorithms, for example, VMF and DDF.
Vector algorithm, in the applications of multi-channel color image processing, has occupied a very important position. However, based on different actual requirements, different algorithms are needed to different fields. For the first category algorithm, it simply choose the common window, and then detect the noisy point, the results have proved that, the result is better, its objective indicators have improved significantly, but this improved algorithm has the different the effect to the different images, which have different characteristics. And the phenomenons of noisy point-leakage exist, which is a flaw of this algorithm. The second improved algorithm is improved algorithm on the base of the first algorithm that is the analysis of the first algorithms. But in the future, the applications of the vector algorithms yet to be developed, but the effect is not very obvious, this also needs to continue to improve in the future.
The simulation results show that the use of the improved noise-detection operator after processing, the effect of image noise removal improved, and maintain a very good details and color preservation. At the same time it could be seen that, the difference of three Operators of the first type, is not great and the choice could be according to the practical needs. Moreover, the use of the Gaussian window in the DDF or VMF, has gained the better effects than the classical VMF and DDF, and has the better property of details preservation and color preservation. And the second operator, in order to use their respective advantages of VMF and DDF, distinguish between noise in the image area and details.
In short, because the images can be easily polluted by the noises of the wireless systems and the noises have the serious impact on the quality of the image. Therefore, removal of the noise in the image to the image processing is a very important part, and, noise suppressing, to the feature-extraction, target- identification, region-segmentation is a very important premise. Noise-suppressing in the video, is also on the base of image processing.
Vector color image filtering is a new development area. Image noise-suppressing and correction is of very important significance, and they have now become emerging in the field, such as medical images, geological images, cultural heritage reservations, video communications, image post-processing machines, objective- identification, and so on. For the further, the vector image filtering technologies have opened a world of limitless possibilities.
Color image processing technology in the present image science, communications, and multimedia applications has a very important position.
本文针对彩色图像脉冲噪声的去除算法进行了研究。根据彩色图像的特性,选择了结合了矢量中值滤波器(VMF)和矢量方向滤波器(VDF)的优点的方向-距离滤波器(DDF)作为基础。并提出了两种改进的噪声去除算法。第一种算法首先进行了区域检测,然后使用第一种噪声检测算子,最后使用高斯加权窗的DDF算法,处理被噪声污染的点;第二种算法是在第一种算法的基础上改进的,它的特点是运算量比第一种改进算法小,而且使用第二种检测算子,将疑为噪声点分为孤立噪声点和细节区,并利用VMF和DDF的优点,对上述区域分别使用VMF和高斯加权窗的DDF算法。
仿真结果表明,使用了第一种改进的噪声去除算法的处理之后,图像的噪声去除效果较好,而且细节和色度保持很好。而且,使用了高斯加权窗的DDF方法比经典的VMF和VDF有更好的细节保持效果和较好的色度保持效果;而第二种改进算法,相比第一种改进算法是一种快速检测算法,目的是为了在图像中区分孤立噪声点和细节区,并利用VMF和DDF各自的优点。实验结果证明,第二种改进算法也具有较好的噪声去除效果。
The color image processing system has been employed in many fields, from image-capture, to feature-abstraction, and the most common processing technique is filtering. And for the most cases, the noises are additive. Filtering is of the most importance. While during the filtering techniques, the most common one is the linear filter, and it is well-known for the simple mathematic expression, the uniform theory basics and simple implementation. However, most of them are on the base of Gauss statistics model.
But, many questions could not been solved by the linear filtering technology. For example, the linear technologies have been limited by the application of image processing. Traditional linear technologies could not become the model of image-forming or image-transportation, and it has not considered the nonlinear model of human vision system. The images compose of smooth fields and the detailed fields, especially, the details mean the better to vision system. The filters which have the characteristic of detail-preservation have gained more and more applications. But most of the linear processing systems will make the details and the borderlines vague.
The noises will degrade the comprehension of human to images, so it becomes the limited factor. Similarly, if you put the images to the quantitative analysis, the noises will restrict the performance of the systems. As a result, noises compression has been the key part of the image processing system, which restores the original images from the images polluted by the impulsive noises.
There are noises in the process of images forming, transportation, and the decoding. The noises make the images color distortion, definition degradation. Due to the different reasons of generation, there are several types of noises. And the most generic one is the impulsive noise, and its impact to images is not negligent. Transportation errors, random movements of camera, instability of electrons, electromagnetic disturbance, errors of sensors, it is common in the transportation in the airs, including the light, the mechanical movements, the motor engines, or high-voltage line errors, and all kinds of electronic switches. And these become the interferences to the broadcast and the communication signals, while the corresponding images will be noisy. But the huge errors will generate a kind of impulsive noises, which appears as black or white pot, so it is called Peppers and Salt noises.
The median filter is well-known for its effects of noise suppression to impulsive noise, but stretches the median filter to the color images simply, that is to imply the median to the RGB channel separately, will not generate good effects, since there are relations between the three channels. So the color distortion will appear. Due to these reasons, the vector median filters appear.
Because the shortcomings of general vector algorithms are: they consider the threads and details as noises, then to process as usual. The image quality will decline, too. Therefore, noise-detection algorithm is to identify the noises, fine details, at the same time, separate the marginal fields and the details from the points which were noise polluted. That is the aim of noise-detection. But the most important thing is to distinguish between lines or the details and the noisy points, so, we introduced the edge-detection algorithms, thereby, to suppress the noises. This paper summarizes the main representative algorithms of several edge- detection algorithms and noise-detection algorithms, firstly, introduced some common edge-detection algorithm, and then introduced several representative algorithms, the first is CANNY edge-detection operator, which is a kind of gradient algorithm operator, the second one is the CIELAB color space noise-detection algorithm.
To the gray images, the edges could be seen as the points whose gray scale jump, or they can be described as the gradual change, or the direction of the maximum gradient, in the gray-scale images, gradient operator can be used to detect image edges, but it is effective to the step edge only. When the gray width increases, the second derivative is a good method, instead, for example, the Laplace operator. And the edges could be described as follows: the zero crossing point of the second derivative. However, in the color image processing, color images are of the multi-channel, so vector processing method is more applicable.
Therefore, in this paper, two types improved noise detection operator have been proposed. The first type is based on the vector computing method, and the second kind is a rapid detection, after the detection, the images were dealt with different methods. Firstly, VMF and DDF filter have been chosen as a basis, but in order to choose a common filter window, so the simulation analysis has been implemented. In this paper, improved color image filtering algorithms have been divided into two types: the first type algorithm, the image of the first divided into different districts, then imply the first noises detection operator to detect the noises points, finally, to used the improved filtering window to the DDF, or VMF; and for the second type algorithm, it is the fast detection algorithm, and it deal with the noisy point and the details with different methods. In the first type algorithm, it is on the basis of the first improved noise-detection algorithms, because it has compared the pros and cons of VMF and DDF, therefore, in the second type of algorithm, firstly, to use the rapid detection, which detected the noisy points, at the same time to avoid unnecessary treatment, such as the threads, then the noisy points could be detected and divided into the isolated noisy points and the details fields, and finally treat these noisy points with different algorithms, for example, VMF and DDF.
Vector algorithm, in the applications of multi-channel color image processing, has occupied a very important position. However, based on different actual requirements, different algorithms are needed to different fields. For the first category algorithm, it simply choose the common window, and then detect the noisy point, the results have proved that, the result is better, its objective indicators have improved significantly, but this improved algorithm has the different the effect to the different images, which have different characteristics. And the phenomenons of noisy point-leakage exist, which is a flaw of this algorithm. The second improved algorithm is improved algorithm on the base of the first algorithm that is the analysis of the first algorithms. But in the future, the applications of the vector algorithms yet to be developed, but the effect is not very obvious, this also needs to continue to improve in the future.
The simulation results show that the use of the improved noise-detection operator after processing, the effect of image noise removal improved, and maintain a very good details and color preservation. At the same time it could be seen that, the difference of three Operators of the first type, is not great and the choice could be according to the practical needs. Moreover, the use of the Gaussian window in the DDF or VMF, has gained the better effects than the classical VMF and DDF, and has the better property of details preservation and color preservation. And the second operator, in order to use their respective advantages of VMF and DDF, distinguish between noise in the image area and details.
In short, because the images can be easily polluted by the noises of the wireless systems and the noises have the serious impact on the quality of the image. Therefore, removal of the noise in the image to the image processing is a very important part, and, noise suppressing, to the feature-extraction, target- identification, region-segmentation is a very important premise. Noise-suppressing in the video, is also on the base of image processing.
Vector color image filtering is a new development area. Image noise-suppressing and correction is of very important significance, and they have now become emerging in the field, such as medical images, geological images, cultural heritage reservations, video communications, image post-processing machines, objective- identification, and so on. For the further, the vector image filtering technologies have opened a world of limitless possibilities.
Color image processing technology in the present image science, communications, and multimedia applications has a very important position.
引文
[1]. Jaakko Astola, member IEEE, Petri Havisto, student member, IEEE and Yrjo Neuvo, fellow, IEEE,Vector Median Filters,Proceedings of IEEE, Vol,78, No.4, April 1990 pp678-688
[2]. 刘丽梅,孙玉荣,李莉,中值滤波技术发展研究,云南师范大学学报,2004.1,24(1):23-27
[3]. Rastislav Lukac, Bogdan Smolka,karl Martin, Konstantions N.Plataniotis, and Anastasios N.Venetsanopulous,Vector Filtering for Color Imaging,IEEE Signal Processing Magazine,January 2005, pp74-83
[4]. P.E.Trahanias and A.N. Venetsanopoulos,Vector Directional Filters-A New Class of Multichannel Image Processing Filters,IEEE Transactions on Image Processing, Vol.2, No.4.October 1993, pp528-534
[5]. Panos E,Trahanias, member, IEEE, Damianos Karakos, and Anastasios N.menetsanopoulos, fellow, IEEE, Directional Processing of Color Images: Theory and Experimental Results, IEEE Transactions on Image Processing, Vol.5, No.6, June 1996, pp868-880
[6]. K.N.Plataniotis,D.and Routsos, and A.N.Venetsanopoulos, Color Image Precessing Using Adaptive Vectoer Directional Filters,IEEE, Transactions on Circuits and Systems-Analog and Digital Signal ,Processing ,Vol.45, No.10, October 1998, pp1414-1419
[7]. Jaura E.Jackon, George N. Rouskas, Matthias F.M.Stallman,The Directional p-median Problem: Definition, Complexity, and Algorithms , Elsevier European Journal of Operational Rescarch Vol.179 (2007), pp1097-1108
[8]. 0-8186-7310-9/95,Damianos G.Karakos and Panos E.trahanias, Combining Vector Median and Vector Directional Filters: the Directional-Distance Filters IEEE, 1995, pp171-174
[9]. Damiamos G.Karakos and Panos E.Trahanias,Generalized Multichannel Image-Filtering Structures,IEEE Transactions on Image Processing, Vol.6, No.7. July 1997, pp868-880
[10]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo, fellow, IEEE, Nonlinear Multivariate Image Filtering Techniques,IEEE Transactions on Image Processing, Vol.4, No.6, June 1995, pp241-244
[11]. Russel C, Hardie and Gonzalo R.arce, member, IEEE,Ranking in R k and its Use in Multivariate Image Estimation,IEEE Transactions on Circuits and Systems for Video Technology, Vol.1, No.2 June 1991, pp197-209
[12]. K.N.Plataniotis, Vinayagamoorthy Sri, D.and Routsos, and A.N.Venetsanopoulos, An Adaptive Nearest Multichannel Filter, IEEE Transactions on Circuirs and Systems forVideo Technology, Vol.6, No.6, December 1996, pp699-703
[13]. K.N.Plataniotis, D.Androutsos, and A.N.Venetsanopoulos, Color Image Processing Using Adaptive Vector Directional Filters, IEEE Transactions on Circuits and Systems, Analog and Digital Signal Processing ,Vol.45, No.10, October,1998, pp1414-1419
[14]. Sung-Jea Ko, member, IEEE and Yong Hoon IEEE, member IEEE,Center Weighted Median Filters and Their Applications to Image Enhancement, IEEE Transactions on Circurts and Systems, Vol, 38, No.9, September 1991, pp984-992
[15]. Sudhakar Kalluri, student member, IEEE, and Gonzalo R.Arce, senior member, IEEE,Adaptive Weighted Myriad Filter Algorithms for Robust Signal Processing in α -stable Noise Environments,IEEE Transactions on Signal Processing, Vol.46, No.2, February 1998, pp322-334
[16]. Laurent Lucata,1, Pierre Siohanb,1, Dominique Barba,Adaptive and Global Optimization Methods for Weighted Vector Median Filters,Signal Processing: Image Communication Vol.17 (2002), pp509–524
[17]. Rastislav Lukac,a,* Bogdan Smolka,b, Konstantinos N. Plataniotis,a and,Anastasios N. Venetsanopoulosa,Selection weighted vector directional filters,Computer Vision and Image Understanding, Computer Vision and Image Understanding Vol.94, (2004), pp140–167
[18]. Tz-Chao Lin,A New Adaptive Center Weighted Median Filter for Suppressing Impulsive Noise in Images,Elsevier, Information Sciences (2006), pp1-15
[19]. 王明常,邢立新,杨毅恒,刘志会,动态方向加权二维多级中值滤波的图像处理,吉林大学学报(信息科学版),2006.1,24(1):18-21
[20]. 0-7803-5471-0/99, Lazhar Khriji, Moncef gabbouj, Vector Median-Rational Hybrid Filters for Multichannel Image Processing, 1999 IEEE, pp90-93
[21]. 张旭明,徐滨士,董世运,甘小明,自适应中值-加权均值混合滤波器,光学技术, 2004.11,30(6):652-656
[22]. Zhonghua Ma, member, IEEE, Hong Ren Wu, and Bin Qiu, senior member, IEEE,A Robust Structure-adaptive Hybrid Vector Filter for Color Image Restoration,IEEE Transactions on Image Processing , Vol.14, No.12, December 2005, pp1990-2001
[23]. P.Pankajakan, V.Kumar ,Detail_Preserving Image Information Restoration Guided by SVM Based Noise Mapping, Elsevier Digital Signal Processing 2006, pp1-17
[24]. Yrjo Neuvo and Walter H. Ku, member, IEEE,Analysis and Digital Realization of a pseudorandom Gaussian and impulsive Noise Source, IEEE Transactions on communications, Vol.com-23, No.9, September 1975, pp850-859
[25]. Michael Vrhel, Eli Saber, and H.Joel Trussell, Color Image Generation and Display Technologies, an Overview of Methods, Devices, and Research, IEEE Signal Processing Magezinge January 2005, pp23-33
[26]. Rajeev Ramanath, Wesley E. Snyder, Youngjun Yoo, and Mark S. Drew,Color ImageProcessing Pipeline, A General Survey of Digital Still Camera Processing,IEEE Signal Processing Magazine January 2005, pp34-43
[27]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo, fellow, IEEE, Multichannel Edge Enhancement in Color Image Processing,IEEE Transactions on Circuits and Systems for Video Technology , Vol ,4, No.5, October, 1994, pp468-479
[28]. Andreas Koschan and Mongi Abidi,Detection and Classification of Edges in Color Images A Review of Vector-Valued Techniques January 2005,IEEE Signal Pressing Magazine, pp64-73
[29]. 朱其刚,具有细节保护特性的多级中值滤波算法,山东科技大学学报(自然科学版),2005.9,24(3):73-75
[30]. 洪林,彭玉华,郭锐,基于方向的自适应多级中值滤波,通信学报,2006.4,27(4):119-123
[31]. How-Lung Eng, student member IEEE, and Kai-Kuang Ma,s enior member IEEE,Noise Adaptive Soft-switching Median Filter,IEEE Transaction on Image Processing, Vol.10, No.2, February 2001, pp242-251
[32]. Bogdan Smolka, and Rzej Chydzinski, Fast Detection and Impulsive Noise Removal in Color Images,Real_time Imaging Vol.11 (2005), pp389-402
[33]. Shih-Chang Hsia,A fast Efficient Restoration Algorithm for High-Noise Image Filtering with Adaptive Approach,J.Vis.Commun.Image.16(2005), pp379-392
[34]. Bogdan Smolka,and Rzej Chydzinski, Fast Detection and Impulsive Noise Removal in Color Images, Reai-Time Imaging Vol.11 (2005), pp389-402
[35]. P.Pankajakan, V, Kumar, Detail-Preserving Image Information Restoration Guided by SVM Based Noise Mapping,Elsevier Digital Signal Processing 2006, pp1-17;
[36]. Stefan Schulte, Valerie De Witte, Mike Nachtegael, Diethich Van Der Weken,Fuzzy Random Impulse Noise Reduction Method, Elsevier, Fuzzy Sets Systems, 2006, pp1-14
[37]. Lianghai Jin, Dehua Li,A switching Vector Median Filter Based on the CIELAB Color Space for Color Image Restoration, Elsevier Signal Processing, 2006, pp1-10
[38]. Dung Dang, Wenbin Luo Engineering Department, St. Mary’s University, San Antonio, TX 78228-8534, USA, Color Image Noise Removal Algorithm Utilizing Hybrid Vector Filtering,Elsevier,int.J.Electron Commun. Vol.62, (2008) pp63-67
[39]. 张旭明,徐滨士,董世运,甘小明,自适应中值-加权均值混合滤波器,光学技术,2004.11,30(6):652-656
[40]. 张清菊,叶明,基于邻域统计和脉冲检测中值滤波改进算法,杭州电子科技大学学报,2005.2,25(1):79-81
[41]. 张旭明,徐滨士,董世运,用于图像处理的自适应中值滤波,计算机辅助设计与图形学学报,2005.2,17(2):295-299
[42]. 许卫全,张新楼。徐中佑,张砚沉,保持图像细节的局部自适应去噪滤波器新算法,信号处理,2005 ,21(2):191-194
[43]. 赵甘露,陈自力,梁四洋,基于改进噪声检测的迭代中值滤波算法,军械工程学院学报,2005.10,17(5):65-67
[44]. 张旭明,徐滨士,董世运,吴毅雄,去除脉冲噪声的自适应开关中值滤波,光电工程,2006.6,33(6):78-83
[45]. Dimitrios and Routsos, studengt member, IEEE, and Anastasios N.Venetsanopoulos, fellow, IEEE , Adaptive Fuzzy Systems for Multichannel Signal Processing ,Konstatntinos N.Plataniotis, member, IEEE, Proceedings of the IEEE, Vol.87. No.9, September 1999 , pp1601-1622
[46]. Stefan Schulte, Valerie De Witte, Mike Nachtegael, Dietrich Van Eer Weken, Etienne E,Kerre, Fuzzy Random Impulse Noise Reduction Method, Elsevier Fuzzy Sets and Systems, 2006, pp1-14
[47]. Joan-Gerard Camarena, Valent?′n Gregori 1, Samuel Morillas *,2, Almanzor Sapena,Fast Detection and Removal of Impulsive Noise Using Peer Groups and Fuzzy Metrics J.vis.commun,Image R .Vol.19 (2008), pp20-29
[48]. Samuel Morillas_, Valent?′n Gregori, Guillermo Peris-Fajarne′ s, Almanzor Sapena,Local Self-Adaptive Fuzzy Filter for Impulsive Noise Removal in color images,Elsevier,signal processing Vol.88 (2008) pp390-398
[49]. Shu-Mei Guo, Chang-Shing IEEE, Chin-Yuan Hsu,An Intelligent Image Agent Based on Soft-computing Techniques for Color Image Processing , Expert Systems with Applications 28(2005), pp483-494;
[50]. M.I, Vardavoulia, I.Aandreadis, PH.Tsalides,A New Vector Median Filter For Color Image Processing, Elsevier Science Pattern Recognition Letters Vol.22 (2001),. pp675-689
[51]. Hao Zhou, K.Z.Mao, An Impulsive Noise Color Image Filter Using Learning-based Color Morphological Operations, Elsevier Digital Signal Processing,2007, pp1-16
[52]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo fellow IEEE, Nonlinear Multivariate Image Filtering Techniques,IEEE Transactions on Image Processing , Vol, 4, No, 6, 1995, pp788-798
[53]. Mauro Barni, Franco Bartolini, and Vito Cappellini,Image Processing for Virtual Restoration of Artworks,University of Florence, Italy Tools for Building Virtual Heritage,IEEE MultiMedia, 2000, 7(2), pp34-37
[54]. 王建卫,彩色图像的中值滤波算法的改进与应用,哈尔滨商业大学学报(自然科学版) 2006.8,22(4):67-69
[55]. Rastislav Lukac, Bogdan Smolka,Konstantinos N.Plataniotis,Sharpening Vector Median Filters, Elsevier, signal processing 87(2007), pp2085-2099
[56]. 郁伯康,郁梅,彩色图像脉冲噪声的自适应矢量滤波,电子科技大学学报,2000.4,29(2):131-135
[57]. 张志龙,曹承倜,李彦青,胡樱,一种保护彩色图像色度信息的矢量中值滤波器,计算机工程与应用,2001 (21) :99-101
[58]. 李永峰,丁润涛,国澄明,一种彩色图像滤波的改进矢量中值滤波算法,数据采集与处理,2003.9,18(3):361-364
[2]. 刘丽梅,孙玉荣,李莉,中值滤波技术发展研究,云南师范大学学报,2004.1,24(1):23-27
[3]. Rastislav Lukac, Bogdan Smolka,karl Martin, Konstantions N.Plataniotis, and Anastasios N.Venetsanopulous,Vector Filtering for Color Imaging,IEEE Signal Processing Magazine,January 2005, pp74-83
[4]. P.E.Trahanias and A.N. Venetsanopoulos,Vector Directional Filters-A New Class of Multichannel Image Processing Filters,IEEE Transactions on Image Processing, Vol.2, No.4.October 1993, pp528-534
[5]. Panos E,Trahanias, member, IEEE, Damianos Karakos, and Anastasios N.menetsanopoulos, fellow, IEEE, Directional Processing of Color Images: Theory and Experimental Results, IEEE Transactions on Image Processing, Vol.5, No.6, June 1996, pp868-880
[6]. K.N.Plataniotis,D.and Routsos, and A.N.Venetsanopoulos, Color Image Precessing Using Adaptive Vectoer Directional Filters,IEEE, Transactions on Circuits and Systems-Analog and Digital Signal ,Processing ,Vol.45, No.10, October 1998, pp1414-1419
[7]. Jaura E.Jackon, George N. Rouskas, Matthias F.M.Stallman,The Directional p-median Problem: Definition, Complexity, and Algorithms , Elsevier European Journal of Operational Rescarch Vol.179 (2007), pp1097-1108
[8]. 0-8186-7310-9/95,Damianos G.Karakos and Panos E.trahanias, Combining Vector Median and Vector Directional Filters: the Directional-Distance Filters IEEE, 1995, pp171-174
[9]. Damiamos G.Karakos and Panos E.Trahanias,Generalized Multichannel Image-Filtering Structures,IEEE Transactions on Image Processing, Vol.6, No.7. July 1997, pp868-880
[10]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo, fellow, IEEE, Nonlinear Multivariate Image Filtering Techniques,IEEE Transactions on Image Processing, Vol.4, No.6, June 1995, pp241-244
[11]. Russel C, Hardie and Gonzalo R.arce, member, IEEE,Ranking in R k and its Use in Multivariate Image Estimation,IEEE Transactions on Circuits and Systems for Video Technology, Vol.1, No.2 June 1991, pp197-209
[12]. K.N.Plataniotis, Vinayagamoorthy Sri, D.and Routsos, and A.N.Venetsanopoulos, An Adaptive Nearest Multichannel Filter, IEEE Transactions on Circuirs and Systems forVideo Technology, Vol.6, No.6, December 1996, pp699-703
[13]. K.N.Plataniotis, D.Androutsos, and A.N.Venetsanopoulos, Color Image Processing Using Adaptive Vector Directional Filters, IEEE Transactions on Circuits and Systems, Analog and Digital Signal Processing ,Vol.45, No.10, October,1998, pp1414-1419
[14]. Sung-Jea Ko, member, IEEE and Yong Hoon IEEE, member IEEE,Center Weighted Median Filters and Their Applications to Image Enhancement, IEEE Transactions on Circurts and Systems, Vol, 38, No.9, September 1991, pp984-992
[15]. Sudhakar Kalluri, student member, IEEE, and Gonzalo R.Arce, senior member, IEEE,Adaptive Weighted Myriad Filter Algorithms for Robust Signal Processing in α -stable Noise Environments,IEEE Transactions on Signal Processing, Vol.46, No.2, February 1998, pp322-334
[16]. Laurent Lucata,1, Pierre Siohanb,1, Dominique Barba,Adaptive and Global Optimization Methods for Weighted Vector Median Filters,Signal Processing: Image Communication Vol.17 (2002), pp509–524
[17]. Rastislav Lukac,a,* Bogdan Smolka,b, Konstantinos N. Plataniotis,a and,Anastasios N. Venetsanopoulosa,Selection weighted vector directional filters,Computer Vision and Image Understanding, Computer Vision and Image Understanding Vol.94, (2004), pp140–167
[18]. Tz-Chao Lin,A New Adaptive Center Weighted Median Filter for Suppressing Impulsive Noise in Images,Elsevier, Information Sciences (2006), pp1-15
[19]. 王明常,邢立新,杨毅恒,刘志会,动态方向加权二维多级中值滤波的图像处理,吉林大学学报(信息科学版),2006.1,24(1):18-21
[20]. 0-7803-5471-0/99, Lazhar Khriji, Moncef gabbouj, Vector Median-Rational Hybrid Filters for Multichannel Image Processing, 1999 IEEE, pp90-93
[21]. 张旭明,徐滨士,董世运,甘小明,自适应中值-加权均值混合滤波器,光学技术, 2004.11,30(6):652-656
[22]. Zhonghua Ma, member, IEEE, Hong Ren Wu, and Bin Qiu, senior member, IEEE,A Robust Structure-adaptive Hybrid Vector Filter for Color Image Restoration,IEEE Transactions on Image Processing , Vol.14, No.12, December 2005, pp1990-2001
[23]. P.Pankajakan, V.Kumar ,Detail_Preserving Image Information Restoration Guided by SVM Based Noise Mapping, Elsevier Digital Signal Processing 2006, pp1-17
[24]. Yrjo Neuvo and Walter H. Ku, member, IEEE,Analysis and Digital Realization of a pseudorandom Gaussian and impulsive Noise Source, IEEE Transactions on communications, Vol.com-23, No.9, September 1975, pp850-859
[25]. Michael Vrhel, Eli Saber, and H.Joel Trussell, Color Image Generation and Display Technologies, an Overview of Methods, Devices, and Research, IEEE Signal Processing Magezinge January 2005, pp23-33
[26]. Rajeev Ramanath, Wesley E. Snyder, Youngjun Yoo, and Mark S. Drew,Color ImageProcessing Pipeline, A General Survey of Digital Still Camera Processing,IEEE Signal Processing Magazine January 2005, pp34-43
[27]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo, fellow, IEEE, Multichannel Edge Enhancement in Color Image Processing,IEEE Transactions on Circuits and Systems for Video Technology , Vol ,4, No.5, October, 1994, pp468-479
[28]. Andreas Koschan and Mongi Abidi,Detection and Classification of Edges in Color Images A Review of Vector-Valued Techniques January 2005,IEEE Signal Pressing Magazine, pp64-73
[29]. 朱其刚,具有细节保护特性的多级中值滤波算法,山东科技大学学报(自然科学版),2005.9,24(3):73-75
[30]. 洪林,彭玉华,郭锐,基于方向的自适应多级中值滤波,通信学报,2006.4,27(4):119-123
[31]. How-Lung Eng, student member IEEE, and Kai-Kuang Ma,s enior member IEEE,Noise Adaptive Soft-switching Median Filter,IEEE Transaction on Image Processing, Vol.10, No.2, February 2001, pp242-251
[32]. Bogdan Smolka, and Rzej Chydzinski, Fast Detection and Impulsive Noise Removal in Color Images,Real_time Imaging Vol.11 (2005), pp389-402
[33]. Shih-Chang Hsia,A fast Efficient Restoration Algorithm for High-Noise Image Filtering with Adaptive Approach,J.Vis.Commun.Image.16(2005), pp379-392
[34]. Bogdan Smolka,and Rzej Chydzinski, Fast Detection and Impulsive Noise Removal in Color Images, Reai-Time Imaging Vol.11 (2005), pp389-402
[35]. P.Pankajakan, V, Kumar, Detail-Preserving Image Information Restoration Guided by SVM Based Noise Mapping,Elsevier Digital Signal Processing 2006, pp1-17;
[36]. Stefan Schulte, Valerie De Witte, Mike Nachtegael, Diethich Van Der Weken,Fuzzy Random Impulse Noise Reduction Method, Elsevier, Fuzzy Sets Systems, 2006, pp1-14
[37]. Lianghai Jin, Dehua Li,A switching Vector Median Filter Based on the CIELAB Color Space for Color Image Restoration, Elsevier Signal Processing, 2006, pp1-10
[38]. Dung Dang, Wenbin Luo Engineering Department, St. Mary’s University, San Antonio, TX 78228-8534, USA, Color Image Noise Removal Algorithm Utilizing Hybrid Vector Filtering,Elsevier,int.J.Electron Commun. Vol.62, (2008) pp63-67
[39]. 张旭明,徐滨士,董世运,甘小明,自适应中值-加权均值混合滤波器,光学技术,2004.11,30(6):652-656
[40]. 张清菊,叶明,基于邻域统计和脉冲检测中值滤波改进算法,杭州电子科技大学学报,2005.2,25(1):79-81
[41]. 张旭明,徐滨士,董世运,用于图像处理的自适应中值滤波,计算机辅助设计与图形学学报,2005.2,17(2):295-299
[42]. 许卫全,张新楼。徐中佑,张砚沉,保持图像细节的局部自适应去噪滤波器新算法,信号处理,2005 ,21(2):191-194
[43]. 赵甘露,陈自力,梁四洋,基于改进噪声检测的迭代中值滤波算法,军械工程学院学报,2005.10,17(5):65-67
[44]. 张旭明,徐滨士,董世运,吴毅雄,去除脉冲噪声的自适应开关中值滤波,光电工程,2006.6,33(6):78-83
[45]. Dimitrios and Routsos, studengt member, IEEE, and Anastasios N.Venetsanopoulos, fellow, IEEE , Adaptive Fuzzy Systems for Multichannel Signal Processing ,Konstatntinos N.Plataniotis, member, IEEE, Proceedings of the IEEE, Vol.87. No.9, September 1999 , pp1601-1622
[46]. Stefan Schulte, Valerie De Witte, Mike Nachtegael, Dietrich Van Eer Weken, Etienne E,Kerre, Fuzzy Random Impulse Noise Reduction Method, Elsevier Fuzzy Sets and Systems, 2006, pp1-14
[47]. Joan-Gerard Camarena, Valent?′n Gregori 1, Samuel Morillas *,2, Almanzor Sapena,Fast Detection and Removal of Impulsive Noise Using Peer Groups and Fuzzy Metrics J.vis.commun,Image R .Vol.19 (2008), pp20-29
[48]. Samuel Morillas_, Valent?′n Gregori, Guillermo Peris-Fajarne′ s, Almanzor Sapena,Local Self-Adaptive Fuzzy Filter for Impulsive Noise Removal in color images,Elsevier,signal processing Vol.88 (2008) pp390-398
[49]. Shu-Mei Guo, Chang-Shing IEEE, Chin-Yuan Hsu,An Intelligent Image Agent Based on Soft-computing Techniques for Color Image Processing , Expert Systems with Applications 28(2005), pp483-494;
[50]. M.I, Vardavoulia, I.Aandreadis, PH.Tsalides,A New Vector Median Filter For Color Image Processing, Elsevier Science Pattern Recognition Letters Vol.22 (2001),. pp675-689
[51]. Hao Zhou, K.Z.Mao, An Impulsive Noise Color Image Filter Using Learning-based Color Morphological Operations, Elsevier Digital Signal Processing,2007, pp1-16
[52]. Kaijun Tang, Jaakko Astola, member, IEEE, and Yrjo Neuvo fellow IEEE, Nonlinear Multivariate Image Filtering Techniques,IEEE Transactions on Image Processing , Vol, 4, No, 6, 1995, pp788-798
[53]. Mauro Barni, Franco Bartolini, and Vito Cappellini,Image Processing for Virtual Restoration of Artworks,University of Florence, Italy Tools for Building Virtual Heritage,IEEE MultiMedia, 2000, 7(2), pp34-37
[54]. 王建卫,彩色图像的中值滤波算法的改进与应用,哈尔滨商业大学学报(自然科学版) 2006.8,22(4):67-69
[55]. Rastislav Lukac, Bogdan Smolka,Konstantinos N.Plataniotis,Sharpening Vector Median Filters, Elsevier, signal processing 87(2007), pp2085-2099
[56]. 郁伯康,郁梅,彩色图像脉冲噪声的自适应矢量滤波,电子科技大学学报,2000.4,29(2):131-135
[57]. 张志龙,曹承倜,李彦青,胡樱,一种保护彩色图像色度信息的矢量中值滤波器,计算机工程与应用,2001 (21) :99-101
[58]. 李永峰,丁润涛,国澄明,一种彩色图像滤波的改进矢量中值滤波算法,数据采集与处理,2003.9,18(3):361-364