运动模糊图像恢复算法的研究与实现
|
摘要
运动模糊图像复原是图像复原技术中十分重要的一个分支,在生产生活领域、航天领域、智能交通领域都有着广泛的应用。由于匀速直线运动模糊是具有普遍意义的一种退化方式,本文针对匀速直线运动模糊图像的复原进行了系统的研究。
建立恰当的退化模型和准确的辨识模糊参数是良好复原退化图像的关键。本文首先根据匀速直线运动模糊图像的特点建立了相应的退化模型,得出其点扩散函数是由模糊长度和模糊角度确定的。
对于匀速直线运动模糊图像,其频谱图像中存在平行排布的暗条纹,这些暗条纹的生成与退化图像的模糊参数存在特定的关系。本文通过对匀速运动模糊图像的频谱出现平行暗条纹的原因的分析,推导了匀速运动模糊图像点扩散函数的离散域表达式,找到了退化图像频谱暗条纹方向和间距与退化图像模糊参数之间的关系式。本文使用Radon算法对暗条纹方向进行检测,使用投影法对暗条纹间距进行检测,然后根据得出的关系式确定退化图像的模糊角度和模糊长度。
对于真实退化的运动模糊图像,使用维纳滤波法对其进行复原后,复原图像会出现较强的振铃效应,本文通过分析局部退化图像形成的原理,找到了一种对局部退化图像预处理的方法,使真实局部退化图像逼近仿真退化图像,这种预处理方法可以有效地改善具有简单背景的退化图像的复原质量。
Motion blurred image restoration is an important branch in the digital image processing, which has been widely used in various fields as production and living, spaceflight, and intelligent transportation. The restoration of uniform linear motion blurred image, the degradation mode of which has a universal significance, is discussed in details in this paper.
The key point to restore a degraded image is to establish the degradation model appropriately and identify the blur parameters accurately. Firstly, the degradation model is established according to the features of a uniform linear motion blurred image, which reveals that the point spread function of a degradation model is determined by the motion blur direction and blur length.
Parallel-arranged dark strips, the generation of which has a particular relationship with the motion blur parameters; can be found in the spectrum of a uniform linear motion blurred image. The mechanism of the generation of parallel-arranged dark strips is analyzed in details in this paper, through which the point spread function (PSF) and the 2-d discrete expression of degraded images are deduced, and through which the correct relationship between linear motion blur parameters and the orientation and space of the dark stripes are found. The Radon transform and Projection transform are used respectively to detect the orientation of strips and the space between strips, and parameters blur angle and blur length of the degraded image are determined consequently.
Through the analysis of degradation principle of local degraded image, an image preprocessing method for a local degraded image is presented in this paper, through which a practical degraded motion blurred image that usually has significant ringing effect after being restored by using wiener filtering can have a similar quality with stimulated image, and the effectiveness of which has been proved in improving the restoration quality of degraded image with simple background.
建立恰当的退化模型和准确的辨识模糊参数是良好复原退化图像的关键。本文首先根据匀速直线运动模糊图像的特点建立了相应的退化模型,得出其点扩散函数是由模糊长度和模糊角度确定的。
对于匀速直线运动模糊图像,其频谱图像中存在平行排布的暗条纹,这些暗条纹的生成与退化图像的模糊参数存在特定的关系。本文通过对匀速运动模糊图像的频谱出现平行暗条纹的原因的分析,推导了匀速运动模糊图像点扩散函数的离散域表达式,找到了退化图像频谱暗条纹方向和间距与退化图像模糊参数之间的关系式。本文使用Radon算法对暗条纹方向进行检测,使用投影法对暗条纹间距进行检测,然后根据得出的关系式确定退化图像的模糊角度和模糊长度。
对于真实退化的运动模糊图像,使用维纳滤波法对其进行复原后,复原图像会出现较强的振铃效应,本文通过分析局部退化图像形成的原理,找到了一种对局部退化图像预处理的方法,使真实局部退化图像逼近仿真退化图像,这种预处理方法可以有效地改善具有简单背景的退化图像的复原质量。
Motion blurred image restoration is an important branch in the digital image processing, which has been widely used in various fields as production and living, spaceflight, and intelligent transportation. The restoration of uniform linear motion blurred image, the degradation mode of which has a universal significance, is discussed in details in this paper.
The key point to restore a degraded image is to establish the degradation model appropriately and identify the blur parameters accurately. Firstly, the degradation model is established according to the features of a uniform linear motion blurred image, which reveals that the point spread function of a degradation model is determined by the motion blur direction and blur length.
Parallel-arranged dark strips, the generation of which has a particular relationship with the motion blur parameters; can be found in the spectrum of a uniform linear motion blurred image. The mechanism of the generation of parallel-arranged dark strips is analyzed in details in this paper, through which the point spread function (PSF) and the 2-d discrete expression of degraded images are deduced, and through which the correct relationship between linear motion blur parameters and the orientation and space of the dark stripes are found. The Radon transform and Projection transform are used respectively to detect the orientation of strips and the space between strips, and parameters blur angle and blur length of the degraded image are determined consequently.
Through the analysis of degradation principle of local degraded image, an image preprocessing method for a local degraded image is presented in this paper, through which a practical degraded motion blurred image that usually has significant ringing effect after being restored by using wiener filtering can have a similar quality with stimulated image, and the effectiveness of which has been proved in improving the restoration quality of degraded image with simple background.
引文
[1]阮秋琦.数字图像处理学.北京：电子工业出版社.2001
[2]刘政凯,瞿建雄.数字图象恢复与重建.合肥：中国科技大学出版社,1989.1-2,148
[3]Cannon T M, Trussell H J, Hunt B R. Comparison of Image Restoration Methods. Applied Optics,1978,17(21):3384～3390
[4]Yitzhaky Y, Kopeika N S. "Identification of blur parameters from motion blurred images," Graph. Models Image Process.1997,59:310～320
[5]徐立云,李霄峰,张斌,邵惠鹤,张健民,任德祥.基于霍夫变换的模糊建模研究及其应用[J].系统仿真学报,2001,(S1)
[6]黄琦,张国基,唐向京.基于霍夫变换的图像运动模糊角度识别法的改进[J].计算机应用,2008,(S1)
[7]Masayuki Tanaka, Kenichi Yoneji, and Masatoshi Okutomi. Motion Blur Parameter Identification from a Linearly Blurred Image [J] Consumer Electronics, 2007.ICCE 2007.Digest of Technical Papers. International Conference on 10-14 Jan.2007 page(s):1-2
[8]Mohsen Ebrahimi MoghaddamMansour JamzadHamid Reza Mahini. Motion Blur Identification in Noisy Images Using Feed-Forward Back Propagation Neural Network. Advances in Machine Vision, Image Processing, and Pattern Analysis 2006.01.01
[9]贺卫国,黎绍发.匀速直线运动模糊长度的精确估计[J].计算机应用,2005(06)：1317-1320.
[10]马明,运动模糊图像的判定与恢复(硕士学位论文),大连,大连理工大学,2006
[11]Castleman K R. Digital Image Processing. Prentice Hall,1998
[12]Helstrom C W. Image restoration by the method of least squares. Journal of the Optical Society of America,1967,57:297～303
[13]曹毅.基于遗传算法的模糊图像复原方法研究[D].东北师范大学,2004
[14]胡广书.数字信号处理理论、算法与实现.清华大学出版社.2003年
[15]程佩青.数字信号处理教程.清华大学出版社.2002年.
[16]雷超阳 廖海洲 周训斌.采用一种新的滤波方法去除图像噪声.微电子学与计算机.2009年
[17]邹谋炎.反卷积与信号复原.北京：国防工业出版社,2001
[18]Masayuki Tanaka, Kenichi Yoneji, and Masatoshi Okutomi. Motion Blur Parameter Identification from a Linearly Blurred Image [J] Consumer Electronics, 2007.ICCE 2007.Digest of Technical Papers. International Conference on 10-14 Jan.2007 page(s):1-2
[19]周玉,彭召意.运动模糊图像的维纳滤波复原研究[J].计算机工程与应用,2009(19)：181-183.
[20]Xing Yi,Changshui Zhang,Jingdong Wang. Multi-view EM algorithm and its application to color image segmentation.2004 IEEE International Conference on Multimedia and Expo (ICME)
[21]张明谦.运动模糊图像的运动参数估计及复原研究.电子科技大学.2009
[22]Zhou Wang, Alan C. Bovik. A Universal Image Quality Index. IEEE Singal Processing Letters.2002,9
[23]Q. Huynh-Thu and M. Ghanbari. Scope of validity of PSNR in image/video quality assessment. ELECTRONICS LETTERS 19th June 2008 Vol.44 No.13
[24]ZHAN G Renyan, ZHAO Guoling, SU Li. A new edge detection method in image processing [C]. IEEE International Symposium on Communications and In2 formation Technology. Piscataway, NJ, USA:IEEE,2005:4450448.
[25]Xinfu Li, Jiaomin Liu. Edge detection on arc image of low volt2age apparatus [C]. IEEE Proceedings of the Second International Conference on Machine Learning and Cybernetics, Xi'an,2003,2921-2924.
[26]Wenzhao Tan, Jin Zhang. Identification of Motion Blur Direction Based on Analysis of Intentional Restoration Errors. Proceedings of the 2004 IEEE International Conference on Networking, Sensing & Control-Taipei, Taiwan, March 21-23,2004
[27]Navid Norouzi and Mohsen Ebrahimi Moghaddam. Motion Blur Identification Using Image derivative. Signal Processing and Information Technology,2008, Page(s):380-384
[28]王晓红,赵荣椿.任意方向运动模糊的消除[J].中国图象图形学报,2000,5(A6)：525-529.
[29]王晅 肖斌 马建峰.基于Radon和解析Fourier-Mellin变换的尺度与旋转不变目标识别算法.中国图象图形学报,2008年11期
[30]罗军辉,冯平.MATLAB 7.0在图像处理中的应用.机械工业出版社,2006.106～112
[31]陈前荣,陆启生,成礼智等.运动模糊图像点扩散函数尺度鉴别.计算机工程与应用,2004：15-19
[32]贺卫国,黎绍发.匀速直线运动模糊长度的精确估计.计算机应用,2005.Vol.25 No.6Page:1316-1317.1320.
[33]王晓凯,李锋.改进的自适应中值滤波[J].计算机工程与应用,2010,(03)
[34]刘政凯,瞿建雄.数字图像恢复与重建[M].合肥：中国科学技术大学出版社,1989：148-182.
[35]Jansson, P. A. Disconsolation of Images and Spectra [M].NY. Academic Press, 1997.
[36]潘琪.运动模糊仿真图像的正确生成.暨南大学.2005
[37]刘微.运动模糊图像恢复算法的研究与实现.中国科学院研究生院.2005
[38]李宇成,覃妮,王目树.一种新的车牌定位方法及应用[J].计算机工程与应用,2009.6 Vol.45(17),pp148-150
[39]张宪,崔岩.卡尔曼滤波与维纳滤波在运动模糊图像恢复中的应用[J].电脑开发与应用,2008.821(8)
[40]Xing Yu Qi, Li Zhang, Chew Lim Tan. Motion Deblurring for Optical Character Recognition. Proceedings of the 2005 Eight International Conference on Document Analysis and Recognition.
[41]Fu liyan,S.I.Olsen. Restoration of Motion Blurred Images by MDL and Piecewise Constant Image Model.1994 International Symposium on Speech, Image processing and Neural Networks,13-16 April 1994, Hong Kong.
[42]Huei-Yung Lin. Vehicle Speed Detection and Identification from a Single Motion Blurred Image. Proceedings of the Seventh IEEE Workshop on Applications of Computer Vision.
[43]Nikolas P. Galatsanos, Vladimir Z. Mesarovic,etc. Hierarchical Bayesian Image Restoration from Partially Known Blurs. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL.9, NO.10, OCTOBER 2000
[44]Gang Xu, Lei Shi. USING GEODESIC ACTIVE CONTOURS FOR MOTION-BLURRED IMAGES CONTOUR DETECTION. Proceedings of the Seventh International Conference on Machine Learning and Cybernetics, Kunming,12-15 July 2008
[45]Moshe Ben-Ezra and Shree K. Nayar. Motion-Based Motion Deblurring. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL.26, NO.6, JUNE 2004
[2]刘政凯,瞿建雄.数字图象恢复与重建.合肥：中国科技大学出版社,1989.1-2,148
[3]Cannon T M, Trussell H J, Hunt B R. Comparison of Image Restoration Methods. Applied Optics,1978,17(21):3384～3390
[4]Yitzhaky Y, Kopeika N S. "Identification of blur parameters from motion blurred images," Graph. Models Image Process.1997,59:310～320
[5]徐立云,李霄峰,张斌,邵惠鹤,张健民,任德祥.基于霍夫变换的模糊建模研究及其应用[J].系统仿真学报,2001,(S1)
[6]黄琦,张国基,唐向京.基于霍夫变换的图像运动模糊角度识别法的改进[J].计算机应用,2008,(S1)
[7]Masayuki Tanaka, Kenichi Yoneji, and Masatoshi Okutomi. Motion Blur Parameter Identification from a Linearly Blurred Image [J] Consumer Electronics, 2007.ICCE 2007.Digest of Technical Papers. International Conference on 10-14 Jan.2007 page(s):1-2
[8]Mohsen Ebrahimi MoghaddamMansour JamzadHamid Reza Mahini. Motion Blur Identification in Noisy Images Using Feed-Forward Back Propagation Neural Network. Advances in Machine Vision, Image Processing, and Pattern Analysis 2006.01.01
[9]贺卫国,黎绍发.匀速直线运动模糊长度的精确估计[J].计算机应用,2005(06)：1317-1320.
[10]马明,运动模糊图像的判定与恢复(硕士学位论文),大连,大连理工大学,2006
[11]Castleman K R. Digital Image Processing. Prentice Hall,1998
[12]Helstrom C W. Image restoration by the method of least squares. Journal of the Optical Society of America,1967,57:297～303
[13]曹毅.基于遗传算法的模糊图像复原方法研究[D].东北师范大学,2004
[14]胡广书.数字信号处理理论、算法与实现.清华大学出版社.2003年
[15]程佩青.数字信号处理教程.清华大学出版社.2002年.
[16]雷超阳 廖海洲 周训斌.采用一种新的滤波方法去除图像噪声.微电子学与计算机.2009年
[17]邹谋炎.反卷积与信号复原.北京：国防工业出版社,2001
[18]Masayuki Tanaka, Kenichi Yoneji, and Masatoshi Okutomi. Motion Blur Parameter Identification from a Linearly Blurred Image [J] Consumer Electronics, 2007.ICCE 2007.Digest of Technical Papers. International Conference on 10-14 Jan.2007 page(s):1-2
[19]周玉,彭召意.运动模糊图像的维纳滤波复原研究[J].计算机工程与应用,2009(19)：181-183.
[20]Xing Yi,Changshui Zhang,Jingdong Wang. Multi-view EM algorithm and its application to color image segmentation.2004 IEEE International Conference on Multimedia and Expo (ICME)
[21]张明谦.运动模糊图像的运动参数估计及复原研究.电子科技大学.2009
[22]Zhou Wang, Alan C. Bovik. A Universal Image Quality Index. IEEE Singal Processing Letters.2002,9
[23]Q. Huynh-Thu and M. Ghanbari. Scope of validity of PSNR in image/video quality assessment. ELECTRONICS LETTERS 19th June 2008 Vol.44 No.13
[24]ZHAN G Renyan, ZHAO Guoling, SU Li. A new edge detection method in image processing [C]. IEEE International Symposium on Communications and In2 formation Technology. Piscataway, NJ, USA:IEEE,2005:4450448.
[25]Xinfu Li, Jiaomin Liu. Edge detection on arc image of low volt2age apparatus [C]. IEEE Proceedings of the Second International Conference on Machine Learning and Cybernetics, Xi'an,2003,2921-2924.
[26]Wenzhao Tan, Jin Zhang. Identification of Motion Blur Direction Based on Analysis of Intentional Restoration Errors. Proceedings of the 2004 IEEE International Conference on Networking, Sensing & Control-Taipei, Taiwan, March 21-23,2004
[27]Navid Norouzi and Mohsen Ebrahimi Moghaddam. Motion Blur Identification Using Image derivative. Signal Processing and Information Technology,2008, Page(s):380-384
[28]王晓红,赵荣椿.任意方向运动模糊的消除[J].中国图象图形学报,2000,5(A6)：525-529.
[29]王晅 肖斌 马建峰.基于Radon和解析Fourier-Mellin变换的尺度与旋转不变目标识别算法.中国图象图形学报,2008年11期
[30]罗军辉,冯平.MATLAB 7.0在图像处理中的应用.机械工业出版社,2006.106～112
[31]陈前荣,陆启生,成礼智等.运动模糊图像点扩散函数尺度鉴别.计算机工程与应用,2004：15-19
[32]贺卫国,黎绍发.匀速直线运动模糊长度的精确估计.计算机应用,2005.Vol.25 No.6Page:1316-1317.1320.
[33]王晓凯,李锋.改进的自适应中值滤波[J].计算机工程与应用,2010,(03)
[34]刘政凯,瞿建雄.数字图像恢复与重建[M].合肥：中国科学技术大学出版社,1989：148-182.
[35]Jansson, P. A. Disconsolation of Images and Spectra [M].NY. Academic Press, 1997.
[36]潘琪.运动模糊仿真图像的正确生成.暨南大学.2005
[37]刘微.运动模糊图像恢复算法的研究与实现.中国科学院研究生院.2005
[38]李宇成,覃妮,王目树.一种新的车牌定位方法及应用[J].计算机工程与应用,2009.6 Vol.45(17),pp148-150
[39]张宪,崔岩.卡尔曼滤波与维纳滤波在运动模糊图像恢复中的应用[J].电脑开发与应用,2008.821(8)
[40]Xing Yu Qi, Li Zhang, Chew Lim Tan. Motion Deblurring for Optical Character Recognition. Proceedings of the 2005 Eight International Conference on Document Analysis and Recognition.
[41]Fu liyan,S.I.Olsen. Restoration of Motion Blurred Images by MDL and Piecewise Constant Image Model.1994 International Symposium on Speech, Image processing and Neural Networks,13-16 April 1994, Hong Kong.
[42]Huei-Yung Lin. Vehicle Speed Detection and Identification from a Single Motion Blurred Image. Proceedings of the Seventh IEEE Workshop on Applications of Computer Vision.
[43]Nikolas P. Galatsanos, Vladimir Z. Mesarovic,etc. Hierarchical Bayesian Image Restoration from Partially Known Blurs. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL.9, NO.10, OCTOBER 2000
[44]Gang Xu, Lei Shi. USING GEODESIC ACTIVE CONTOURS FOR MOTION-BLURRED IMAGES CONTOUR DETECTION. Proceedings of the Seventh International Conference on Machine Learning and Cybernetics, Kunming,12-15 July 2008
[45]Moshe Ben-Ezra and Shree K. Nayar. Motion-Based Motion Deblurring. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL.26, NO.6, JUNE 2004