运动模糊图像复原算法研究
|
摘要
多媒体技术的迅猛发展以及图像数字化和图像显示设备的普及,为图像处理的发展提供了良好的条件。但图像在获取、传输和存储过程中受各种原因的影响,会造成图像质量的退化。运动模糊图像是由于相机和被拍摄对象之间的相对运动而造成的模糊现象,这一现象在日常生活中经常遇到,因此运动模糊图像复原技术便成为目前图像复原技术的研究热点之一,论文的研究工作正是围绕运动模糊图像复原技术展开。
论文总结了作者在运动模糊图像复原技术中的主要研究工作,包括:分析运动模糊图像的成因以及成像过程;建立运动模糊退化模型;根据传统运动模糊图像复原方法中的不足之处,提出了一种新的方法,降低了原有算法的复杂度;针对传统复原方法易产生振铃效应的问题,分析振铃效应产生的原因,并给出了解决方法;讨论运动模糊参数对于复原过程的重要性,提出了一种新的运动模糊参数鉴别方法。对于上述各种算法,均进行了实验验证和数据分析。
论文对研究过程中取得的主要创新工作进行了详细阐述。这些创新工作简要归纳如下:优化Wiener滤波复原技术过程,在保证图像复原质量的同时,降低了算法运行时间,减少了系统的开销;分析图像边界条件,讨论其对复原结果的影响,通过对图像边界进行处理,在去除图像的模糊和噪声的同时,又保持了图像的细节,改善了复原图像的视觉质量,使图像看起来更加平滑;对比清晰图像与模糊图像的Fourier频谱图,发现Fourier频谱的分布与运动模糊参数有着密切关系,通过检测并分析模糊图像的Fourier频谱图提出了一种新的运动模糊参数鉴别方法,该法提高了模糊参数的鉴别精度;最后为了验证论文中算法的复原效果,采用主客观相结合的方法对复原图像进行质量评价。
With the development of the multimedia technology and the popularization of the digital display device, all of this promotes the development of the digital image processing. Since images are often deteriorated by factors such as blurring and noising during the process of recoding, transmission and storing. Motion-blurred image is the phenomena that the relative motion between moving objects and camera blurs the image. This situation often encountered in daily life, therefore, motion blurred images restoration technology becomes an important hotspot of the image processing recently.
In this dissertation, the author’s major research work is summarized as follows: Analysis of the causes of motion-blurred images and the process of the image; Motion-Blurred image can be modeled; Traditional motion-blurred image restoration methods inadequacies, a new method is proposed to reduce the complexity of the original algorithm; Easy Ways to recover the traditional ringing effects have questions, analyze the causes of ringing effect and give a solution; Discussing the importance of the motion-blur parameters for the restoration process, and propose a new identification method of fuzzy parameters. A simulation platform is constructed to verify the correctness and to analyze the performance of the above algorithms.
The dissertation describes the research works in detail, the principal contributions of the work presented in this thesis are: Optimizing the process of Wiener Filter, in ensuring the quality of image restoration, algorithm reduces the running time and the cost of the system. Analyzing of the image boundary conditions, to discuss its results on the impact of rehabilitation through treatment of the image border, the fuzzy image in the removal and noise while maintaining the details of the images to improve the recovery of the visual image quality, the image seems more smoothing. Images contrast with the blurred image of Fourier frequency spectrum and found that the distribution of Fourier spectrum with the motion blur parameters are closely related, through the detection and analysis of blurred images, a new motion blurred parameter identification are proposed, which improve the accuracy of identification of motion blurred parameters. Finally, a combination of subjective and objective method of image quality evaluation of recovery is used to verify the effectiveness of the recovery algorithm.
论文总结了作者在运动模糊图像复原技术中的主要研究工作,包括:分析运动模糊图像的成因以及成像过程;建立运动模糊退化模型;根据传统运动模糊图像复原方法中的不足之处,提出了一种新的方法,降低了原有算法的复杂度;针对传统复原方法易产生振铃效应的问题,分析振铃效应产生的原因,并给出了解决方法;讨论运动模糊参数对于复原过程的重要性,提出了一种新的运动模糊参数鉴别方法。对于上述各种算法,均进行了实验验证和数据分析。
论文对研究过程中取得的主要创新工作进行了详细阐述。这些创新工作简要归纳如下:优化Wiener滤波复原技术过程,在保证图像复原质量的同时,降低了算法运行时间,减少了系统的开销;分析图像边界条件,讨论其对复原结果的影响,通过对图像边界进行处理,在去除图像的模糊和噪声的同时,又保持了图像的细节,改善了复原图像的视觉质量,使图像看起来更加平滑;对比清晰图像与模糊图像的Fourier频谱图,发现Fourier频谱的分布与运动模糊参数有着密切关系,通过检测并分析模糊图像的Fourier频谱图提出了一种新的运动模糊参数鉴别方法,该法提高了模糊参数的鉴别精度;最后为了验证论文中算法的复原效果,采用主客观相结合的方法对复原图像进行质量评价。
With the development of the multimedia technology and the popularization of the digital display device, all of this promotes the development of the digital image processing. Since images are often deteriorated by factors such as blurring and noising during the process of recoding, transmission and storing. Motion-blurred image is the phenomena that the relative motion between moving objects and camera blurs the image. This situation often encountered in daily life, therefore, motion blurred images restoration technology becomes an important hotspot of the image processing recently.
In this dissertation, the author’s major research work is summarized as follows: Analysis of the causes of motion-blurred images and the process of the image; Motion-Blurred image can be modeled; Traditional motion-blurred image restoration methods inadequacies, a new method is proposed to reduce the complexity of the original algorithm; Easy Ways to recover the traditional ringing effects have questions, analyze the causes of ringing effect and give a solution; Discussing the importance of the motion-blur parameters for the restoration process, and propose a new identification method of fuzzy parameters. A simulation platform is constructed to verify the correctness and to analyze the performance of the above algorithms.
The dissertation describes the research works in detail, the principal contributions of the work presented in this thesis are: Optimizing the process of Wiener Filter, in ensuring the quality of image restoration, algorithm reduces the running time and the cost of the system. Analyzing of the image boundary conditions, to discuss its results on the impact of rehabilitation through treatment of the image border, the fuzzy image in the removal and noise while maintaining the details of the images to improve the recovery of the visual image quality, the image seems more smoothing. Images contrast with the blurred image of Fourier frequency spectrum and found that the distribution of Fourier spectrum with the motion blur parameters are closely related, through the detection and analysis of blurred images, a new motion blurred parameter identification are proposed, which improve the accuracy of identification of motion blurred parameters. Finally, a combination of subjective and objective method of image quality evaluation of recovery is used to verify the effectiveness of the recovery algorithm.
引文
[1] Ozkan M K, Erdem A T, Sezan M I, et al. Efficient multiframe Wiener restoration of blurred and noisy image sequences. IEEE Transaction on Image Processing[J], 1992, 1(4):453~476
[2] Wu W, Kundu A. Image estimation using fast modified reduced update Kalman filter. IEEE Transaction on Signal processing[J], 1992, 40(4):915~926
[3] Citrin S, Azimi-Sadjadi M R. A full-plane block Kalman filters for image restoration. IEEE Transaction on Image processing[J], 1992, 1(4):488~495
[4] Koch S, Kaufman H, Biemond J. Restoration of spatially varying blurres images using multiple model-based extended Kalman filters. IEEE Transactions on Image Processing[J], 1995, 4(4):520~523
[5] Koch S, Kaufman H, Biemond J. Restoration of spatially varying blurres images using multiple model-based extended Kalman filters. IEEE Transactions on Image Processing[J], 1995, 4(4):520~523
[6] Erler K, Jernigan E. Adaptive image restoration using recursive image filters. IEEE Transactions on Signal Processing[J], 1994, 42(7):1877~1881
[7] Mallikarjuna H S, Chaparro L F. Iterative composite filtering for image restoration. IEEE Transactions on pattern Analysis and Machine Intelligence [J]. 1992, 14(6):674~678.
[8] Kunder D, Hatzinakos D. A Novel blind deconvolution scheme for image restoration using recursive filtering. IEEE Transactions on Signal Processing[J], 1998, 48(2):375~390
[9] Chow T W S, Li X D, ChoS Y, Improved blind image restoration scheme using recurrent filtering. IEEE Proceedings-Vision, Image and Signal Processing[J], 2000, 147(1):23~28
[10] Rafael C G, Richared E W, Digital Image Processing. Second Edition [M]. Beijing: Publishing House of Electronics Industry, 2002:231~234
[11] Wendt P D, Coyle E J, Gallagher N C J. Stack filters [J]. IEEE Trans on Acoust, Speech, Signal Processing, 1986, 34(4):898~911.
[12] Serra J. Image analysis and mathematical morphology, New York: Academic Press, 1982:34~40
[13]蔡靖,杨晋生,丁润涛.模糊加权均值滤波器[J].中国图像图形学报,2000, 5(A)(1):52~56
[14]王伟,赵春辉,孙圣和.最优全方位结构元约束二维层叠滤波器[J].中国图像图形学报,1999,4(A)(6):445~449
[15]赵春辉,王伟.全方位多结构元形态滤波器[J].中国图像图形学报,1997,2(7):500~503
[16] Eng H L Ma K K. Noise adaptive soft-switching median filter [J]. IEEE Transaction on Image Processing, 2001, 10(2): 242~251
[17] Gouchol P, Liu J C, Nair A S. Selective removal of impulse noise based in homogeneity levelinformation[J]. IEEE Transaction on image Processing, 2003, 12(1):85~92.
[18]吴斌,吴亚东等.基于变分微分方程的图像复原技术[M].北京大学出版社,2008
[19]赵晓宇,陈刚,李均利等.数字图像处理与UIMP系统的设计与实现[M].浙江大学出版社,2007
[20] Huei-Yung Lin. Vehicle Speed Detection and Identification from a Single Motion Blurred Image, Application of Computer Vision [J], 2005. Seventh IEEE Workshops on Volume 1, Issue, 5-7 Jan. 2005: 461~467
[21] Likas, A.C. Galatsanos, N.P. A Variational Approach for Bayesian Blind Image Deconvolution [J], Signal Processing, IEEE, 2004, Vol52, No.8: 512~516
[22]张秉伊.运动模糊图像的降质过程与恢复技术研究[J].中国图形图像学报,第9卷,第7期,2004:815~819
[23]周维华.Wiener滤波图像复原,计算机工程与科学[J]第29卷,第3期,2007:39~41
[24] Mark R.banham, Aggelos K.Katsaggelos, Digital Image Restoration [J], IEEE Signal Pressing Magazine, March 1997:256~264
[25] Yitzhaky Y,Kopeika N S. Identification of blur parameters from motion blurred images[J].Graphical Models and Image Processing,1997,59(5):310~320.
[26] Kenneth R.Castleman, Digital Image Processing [M],电子工业出版社,北京,2002
[27]陈小蔷,张俊,吴乐南.基于Neumann边界条件的图像局部复原[J].信号处理,2004,20(4):399~402
[28] Likas, A.C. Galatsanos, N.P. A Variational Approach for Bayesian Blind Image Deconvolution [J], Signal Processing, IEEE, 2004, Vol52, No.8: 512~516
[29]蔡汉添,张军萍.一种基于小波变换的迭代正则化图像恢复算法[J].中国图象图形学报,1999,4A(3):229~233
[30] Wang Zhou, Alan C Bovik.A universal image quality index [J]. IEEE Signal Processing Letters, 2002, (9): 81~87
[31] Shohdohji T, Hoshino Y. Optimization of image quality for decoded images using three-dimensional smoothing method [J]. Applied Mathematics and Computation, 2001, 120: 301~309
[32]杨军,张延生.图像质量主客观评价相关性分析[J].军械工程学院学报,2003,15(1):29~33
[33] Wolf S, Voran S, Seitz N, et al. User-Oriented Measures of Telecommunication Quality [J], IEEE Communication Magazine, 1994 32(1): 56~66
[34] Gerald Westtheimer, FERGU SW Campbell. Light Distribution in the Image Formed by the Living Human Eye [J]. Journal of the Optical Society of America, 1962, 52 (9) : 1040~1045
[35] Dou Glasj Granrath. The Role of Human Visual Models in Image Processing [J]. Proceeding of the IEEE, 1981, 69 (5):552~561.
[36] Mannos JL, Sakrison DJ. The effects of a visual fidelity criterion on the encoding of images [J]. IEEETransact ions on Information Theory, 1974, 10 (5) :525~536
[37] Van Dijk AM, Martens JB. Subjective quality assessment of compressed images [J]. Signal Processing, 1997, 58 (4): 235~252.
[38]王晓晖,朱耀庭,朱光喜.解压缩图像质量的客观评价研究[J],中国图像图形学报,1999, 4A (12) : 1029~1032.
[39]张弓,张昆辉,朱兆达等.一种基于逼近信噪比的SAR图像质量评估方法[J],南京航空航天大学学报,2004,36(2):240~244
[40]魏政刚,袁杰辉,蔡元龙.一种基于视觉感知的图像质量评价方法[J].电子学报,1999,27(4):79~82
[41]杨守义,罗伟雄.一种基于高阶统计量的图像质量客观评价方法[J].北京理工大学学报,2001,,21(5) :610~613
[2] Wu W, Kundu A. Image estimation using fast modified reduced update Kalman filter. IEEE Transaction on Signal processing[J], 1992, 40(4):915~926
[3] Citrin S, Azimi-Sadjadi M R. A full-plane block Kalman filters for image restoration. IEEE Transaction on Image processing[J], 1992, 1(4):488~495
[4] Koch S, Kaufman H, Biemond J. Restoration of spatially varying blurres images using multiple model-based extended Kalman filters. IEEE Transactions on Image Processing[J], 1995, 4(4):520~523
[5] Koch S, Kaufman H, Biemond J. Restoration of spatially varying blurres images using multiple model-based extended Kalman filters. IEEE Transactions on Image Processing[J], 1995, 4(4):520~523
[6] Erler K, Jernigan E. Adaptive image restoration using recursive image filters. IEEE Transactions on Signal Processing[J], 1994, 42(7):1877~1881
[7] Mallikarjuna H S, Chaparro L F. Iterative composite filtering for image restoration. IEEE Transactions on pattern Analysis and Machine Intelligence [J]. 1992, 14(6):674~678.
[8] Kunder D, Hatzinakos D. A Novel blind deconvolution scheme for image restoration using recursive filtering. IEEE Transactions on Signal Processing[J], 1998, 48(2):375~390
[9] Chow T W S, Li X D, ChoS Y, Improved blind image restoration scheme using recurrent filtering. IEEE Proceedings-Vision, Image and Signal Processing[J], 2000, 147(1):23~28
[10] Rafael C G, Richared E W, Digital Image Processing. Second Edition [M]. Beijing: Publishing House of Electronics Industry, 2002:231~234
[11] Wendt P D, Coyle E J, Gallagher N C J. Stack filters [J]. IEEE Trans on Acoust, Speech, Signal Processing, 1986, 34(4):898~911.
[12] Serra J. Image analysis and mathematical morphology, New York: Academic Press, 1982:34~40
[13]蔡靖,杨晋生,丁润涛.模糊加权均值滤波器[J].中国图像图形学报,2000, 5(A)(1):52~56
[14]王伟,赵春辉,孙圣和.最优全方位结构元约束二维层叠滤波器[J].中国图像图形学报,1999,4(A)(6):445~449
[15]赵春辉,王伟.全方位多结构元形态滤波器[J].中国图像图形学报,1997,2(7):500~503
[16] Eng H L Ma K K. Noise adaptive soft-switching median filter [J]. IEEE Transaction on Image Processing, 2001, 10(2): 242~251
[17] Gouchol P, Liu J C, Nair A S. Selective removal of impulse noise based in homogeneity levelinformation[J]. IEEE Transaction on image Processing, 2003, 12(1):85~92.
[18]吴斌,吴亚东等.基于变分微分方程的图像复原技术[M].北京大学出版社,2008
[19]赵晓宇,陈刚,李均利等.数字图像处理与UIMP系统的设计与实现[M].浙江大学出版社,2007
[20] Huei-Yung Lin. Vehicle Speed Detection and Identification from a Single Motion Blurred Image, Application of Computer Vision [J], 2005. Seventh IEEE Workshops on Volume 1, Issue, 5-7 Jan. 2005: 461~467
[21] Likas, A.C. Galatsanos, N.P. A Variational Approach for Bayesian Blind Image Deconvolution [J], Signal Processing, IEEE, 2004, Vol52, No.8: 512~516
[22]张秉伊.运动模糊图像的降质过程与恢复技术研究[J].中国图形图像学报,第9卷,第7期,2004:815~819
[23]周维华.Wiener滤波图像复原,计算机工程与科学[J]第29卷,第3期,2007:39~41
[24] Mark R.banham, Aggelos K.Katsaggelos, Digital Image Restoration [J], IEEE Signal Pressing Magazine, March 1997:256~264
[25] Yitzhaky Y,Kopeika N S. Identification of blur parameters from motion blurred images[J].Graphical Models and Image Processing,1997,59(5):310~320.
[26] Kenneth R.Castleman, Digital Image Processing [M],电子工业出版社,北京,2002
[27]陈小蔷,张俊,吴乐南.基于Neumann边界条件的图像局部复原[J].信号处理,2004,20(4):399~402
[28] Likas, A.C. Galatsanos, N.P. A Variational Approach for Bayesian Blind Image Deconvolution [J], Signal Processing, IEEE, 2004, Vol52, No.8: 512~516
[29]蔡汉添,张军萍.一种基于小波变换的迭代正则化图像恢复算法[J].中国图象图形学报,1999,4A(3):229~233
[30] Wang Zhou, Alan C Bovik.A universal image quality index [J]. IEEE Signal Processing Letters, 2002, (9): 81~87
[31] Shohdohji T, Hoshino Y. Optimization of image quality for decoded images using three-dimensional smoothing method [J]. Applied Mathematics and Computation, 2001, 120: 301~309
[32]杨军,张延生.图像质量主客观评价相关性分析[J].军械工程学院学报,2003,15(1):29~33
[33] Wolf S, Voran S, Seitz N, et al. User-Oriented Measures of Telecommunication Quality [J], IEEE Communication Magazine, 1994 32(1): 56~66
[34] Gerald Westtheimer, FERGU SW Campbell. Light Distribution in the Image Formed by the Living Human Eye [J]. Journal of the Optical Society of America, 1962, 52 (9) : 1040~1045
[35] Dou Glasj Granrath. The Role of Human Visual Models in Image Processing [J]. Proceeding of the IEEE, 1981, 69 (5):552~561.
[36] Mannos JL, Sakrison DJ. The effects of a visual fidelity criterion on the encoding of images [J]. IEEETransact ions on Information Theory, 1974, 10 (5) :525~536
[37] Van Dijk AM, Martens JB. Subjective quality assessment of compressed images [J]. Signal Processing, 1997, 58 (4): 235~252.
[38]王晓晖,朱耀庭,朱光喜.解压缩图像质量的客观评价研究[J],中国图像图形学报,1999, 4A (12) : 1029~1032.
[39]张弓,张昆辉,朱兆达等.一种基于逼近信噪比的SAR图像质量评估方法[J],南京航空航天大学学报,2004,36(2):240~244
[40]魏政刚,袁杰辉,蔡元龙.一种基于视觉感知的图像质量评价方法[J].电子学报,1999,27(4):79~82
[41]杨守义,罗伟雄.一种基于高阶统计量的图像质量客观评价方法[J].北京理工大学学报,2001,,21(5) :610~613