运动模糊图像的恢复及恢复质量评价
|
摘要
在图像拍摄记录的过程中,由于被摄物与成像系统产生相对运动而造成的图像退化称为运动模糊。对其复原的目的在于补偿图像降质,使恢复后的图像尽可能地接近原始图像,为此要对恢复图像的质量进行评价。运动模糊参数是建立二维点扩展函数所必须的先验知识,采用基于频域特性,结合高斯-拉普拉斯边缘检测和Radon变换的方法对运动方向进行估计;利用基于波方程和自相关的测量方法鉴别运动模糊距离。在此基础上,采用Richardson-Lucy迭代算法和维纳滤波方法在选取不同参数的情况下对运动模糊图像进行恢复。对于复原的图像,使用基于传播波方程的运动模糊图像恢复质量客观评价准则进行质量评价,通过与均方误差评价指标的比较,验证了该准则的有效性、可靠性及和主观评价结果的一致性,并确定了参数选取对维纳滤波恢复质量的影响。
Motion-blurring is caused by the relative motion between the camera and the object while the camera is recording. The purpose of restoration is to compensate for image degradation, so that the restored image is as close as possible the original image. So the quality of restored image should be evaluated. Motion blur parameters are priori knowledge necessary for building two-dimensional point spread function. The direction of motion blur is estimated by using the method combined Gauss - Laplace edge detection and Radon transform based on frequency-domain features. The measurement method of blur extent identification based on wave equations and autocorrelation function is used. On this basis, the Richardson-Lucy iterative algorithm and Wiener filtering method are utilized for motion-blurred image restoration under different parameters. For the restored images, Objective assessment to restoration of motion-blurred images based on traveling wave equation is adopted to evaluate the images quality. The criterion’s validity, reliability and consistency with subjective evaluation are proven, and the effect of parameters selection of the Wiener filter method to restored image quality is determined through the comparison with the evaluation index of mean square error.
Motion-blurring is caused by the relative motion between the camera and the object while the camera is recording. The purpose of restoration is to compensate for image degradation, so that the restored image is as close as possible the original image. So the quality of restored image should be evaluated. Motion blur parameters are priori knowledge necessary for building two-dimensional point spread function. The direction of motion blur is estimated by using the method combined Gauss - Laplace edge detection and Radon transform based on frequency-domain features. The measurement method of blur extent identification based on wave equations and autocorrelation function is used. On this basis, the Richardson-Lucy iterative algorithm and Wiener filtering method are utilized for motion-blurred image restoration under different parameters. For the restored images, Objective assessment to restoration of motion-blurred images based on traveling wave equation is adopted to evaluate the images quality. The criterion’s validity, reliability and consistency with subjective evaluation are proven, and the effect of parameters selection of the Wiener filter method to restored image quality is determined through the comparison with the evaluation index of mean square error.
引文
[1]沈峘,李舜酩,毛建国等.数字图像复原技术综述.中国图象图形学报. 2009,14(9). 1764-1775
[2]张秉仁,陈里铭,高游.运动模糊图像的降质过程分析与恢复技术研究.中国图象图形学报. 2004,7,9(7). 815 - 819.
[3]张媛.运动模糊图像的处理.安徽大学,2005
[4]黄隆华.基于质量相关因素及结构失真测量的图像质量评价方法.江西财经大学,2006
[5]孟昕.运动模糊图像的处理与恢复研究.安徽大学,2007
[6]周景超,戴汝为,肖柏华.图像质量评价研究综述.计算机科学. 2008,35(7). 1-4
[7]孙艳丽.运动模糊图像的恢复与处理.大连海事大学,2008
[8]谢凤英,赵丹培. Visual C++数字图像处理.北京:电子工业出版社,2008. 239-240,360-365
[9]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(长春光学精密机械与物理研究所),2006年
[10]阮秋琦.数字图像处理学.北京:电子工业出版社,2001. 63-80,305-312.
[11]高梅.运动模糊的点扩展函数及图像恢复研究.华北工学院,2004年
[12]明文华.运动模糊图像复原算法研究.安徽:安徽大学,2004
[13]高梅,陈树越.任意方向直线运动模糊点扩展函数研究.华北工学院学报. 2004,25(4). 297-299
[14]刘晶,李俊,孙涵.改进的Bresenham直线生成算法.计算机应用与软件. 2008,25(10). 247-249
[15]赵剡,杨威.一种基于边缘特征的图像质量评价新方法.红外与激光工程. 2006,10,35(增刊). 232-238
[16] Milan Sonka, Vaclav Hlavac, Roger Boyle著,艾海舟,武勃译.图像处理、分析与机器视觉.北京:人民邮电出版社,2003 [17 ]赵景秀,韩君君,王菁等.一种改进的LoG图像边缘检测方法.计算技术与自动化. 2009,9,28(3). 74-77
[18]严国萍,戴若愚,潘晴等.基于LOG算子的自适应图像边缘检测方法.华中科技大学学报(自然科学版). 2008,3,36(3). 85-87,102
[19] Sifeng Wang, Jingxiu Zhao. A Self-Adaptive Edge Detection Method Based on LoG Algorithm. 2008 International Conference on Audio,Language and Image Processing. 2008.1009-1011.
[20]张晓燕,尤红建,付琨.基于Radon变换的星载SAR图像窄V尾迹检测.测绘科学. 2009,3,34(2). 117-118,111
[21]李秀怡,黄继风.基于Radon变换的运动模糊方向精确估计.计算机工程与科学. 2008,30(9). 51-52,57
[22]周玉,周序生,戴圣伟等.一种改进的运动模糊图像复原算法.计算机工程与科学. 2009,31(10). 49-52
[23]鲁晓东.一种运动模糊图像恢复方法的研究.华中科技大学,2006
[24]张进,陈慧蓉,荣钢.基于波方程的运动模糊图像恢复.清华大学学报(自然科学版). 2005,45(7). 1002-1004,1008
[25] Horn B, Schunck B. Determining optical flow .Artificial Intelligence, 1981,17(1). 185-203.
[26] Rekleitis I. Steerable filters and ceptral analysis for optical flow calculation from a single blurred image. Vision Interface,1996,12(1). 159,166.
[27]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. 2004,40(23). 15-19,49
[28]钱春强,王继成.基于改进约束最小二乘方法的图像复原算法.计算机技术与发展. 2007,17(6). 9-11,14
[29]李红阳.运动图象恢复.清华大学,2000. 18-31
[30]周玉,彭召意.运动模糊图像的维纳滤波复原研究.计算机工程与应用. 2009,45(19). 181-183
[31] M.M.Sondhi. Image Restoration: The Removal of Spatially Invariant Degradations. Proc. IEEE, 1972, 60(7).842-853.
[32]刘艳春.利用传播波方程作运动模糊图像质量恢复的客观评价.广州:暨南大学,2002.
[33]刘明.基于图像复原的航空摄影前向像移检测及补偿技术研究.中国科学院研究生院(长春光学精密机械与物理研究所),2005
[34]杨琬,吴乐华,范晔.数字图像客观质量评价方法研究.通信技术. 2008,7(41). 244-246
[35]倪晓明.基于小波变换图像质量评价新算法.厦门:厦门大学,2009.
[36]路文,高新波,王体胜.一种基于小波分析的部分参考型图像质量评价方法.电子与信息学报. 2009,31(2). 335-338
[37]安雪晶,田媛.无参考图像质量评价方法的设计原则.中国光学与应用光学. 2009,2(2). 140-144
[38] WANG X, TIAN B F, L IANG CH. Blind image quality assessment for measuring image blur. IEEE, 2008 Cong. Image S ig. Proces , Hainan, China, 2008. 467-470.
[39]蔡利栋.运动模糊图像恢复质量的客观评价:传播波方法.北京:第二届全国智能视觉监控学术会议,2003.
[40] Cai Lidong. Traveling Wave Equations and Restoration of Motion-Blurred Images. Acta Automatica Sinica,2003,29(3). 466 - 471.
[41]刘艳春,蔡利栋.基于传播波方程的图像恢复质量客观评价方法的验证.暨南大学学报(自然科学版).2004,25(1). 30-35
[42] Li-Dong Cai.Objective Assessment to Restoration of Global Motion-blurred Images Using Traveling Wave Equations. Multi-Agent Security and Survivability, 2004 IEEE First Symposium on.2004
[43] Cai L D,Zhen J W. Image blur caused by uniform linear motion and its removal using traveling wave equation and Hough transform. PROC of Second INT. CONF. on Image and Graphics,Wei Sui(Ed.),Hefei, China,SPIE, August 2002,268-275
[2]张秉仁,陈里铭,高游.运动模糊图像的降质过程分析与恢复技术研究.中国图象图形学报. 2004,7,9(7). 815 - 819.
[3]张媛.运动模糊图像的处理.安徽大学,2005
[4]黄隆华.基于质量相关因素及结构失真测量的图像质量评价方法.江西财经大学,2006
[5]孟昕.运动模糊图像的处理与恢复研究.安徽大学,2007
[6]周景超,戴汝为,肖柏华.图像质量评价研究综述.计算机科学. 2008,35(7). 1-4
[7]孙艳丽.运动模糊图像的恢复与处理.大连海事大学,2008
[8]谢凤英,赵丹培. Visual C++数字图像处理.北京:电子工业出版社,2008. 239-240,360-365
[9]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(长春光学精密机械与物理研究所),2006年
[10]阮秋琦.数字图像处理学.北京:电子工业出版社,2001. 63-80,305-312.
[11]高梅.运动模糊的点扩展函数及图像恢复研究.华北工学院,2004年
[12]明文华.运动模糊图像复原算法研究.安徽:安徽大学,2004
[13]高梅,陈树越.任意方向直线运动模糊点扩展函数研究.华北工学院学报. 2004,25(4). 297-299
[14]刘晶,李俊,孙涵.改进的Bresenham直线生成算法.计算机应用与软件. 2008,25(10). 247-249
[15]赵剡,杨威.一种基于边缘特征的图像质量评价新方法.红外与激光工程. 2006,10,35(增刊). 232-238
[16] Milan Sonka, Vaclav Hlavac, Roger Boyle著,艾海舟,武勃译.图像处理、分析与机器视觉.北京:人民邮电出版社,2003 [17 ]赵景秀,韩君君,王菁等.一种改进的LoG图像边缘检测方法.计算技术与自动化. 2009,9,28(3). 74-77
[18]严国萍,戴若愚,潘晴等.基于LOG算子的自适应图像边缘检测方法.华中科技大学学报(自然科学版). 2008,3,36(3). 85-87,102
[19] Sifeng Wang, Jingxiu Zhao. A Self-Adaptive Edge Detection Method Based on LoG Algorithm. 2008 International Conference on Audio,Language and Image Processing. 2008.1009-1011.
[20]张晓燕,尤红建,付琨.基于Radon变换的星载SAR图像窄V尾迹检测.测绘科学. 2009,3,34(2). 117-118,111
[21]李秀怡,黄继风.基于Radon变换的运动模糊方向精确估计.计算机工程与科学. 2008,30(9). 51-52,57
[22]周玉,周序生,戴圣伟等.一种改进的运动模糊图像复原算法.计算机工程与科学. 2009,31(10). 49-52
[23]鲁晓东.一种运动模糊图像恢复方法的研究.华中科技大学,2006
[24]张进,陈慧蓉,荣钢.基于波方程的运动模糊图像恢复.清华大学学报(自然科学版). 2005,45(7). 1002-1004,1008
[25] Horn B, Schunck B. Determining optical flow .Artificial Intelligence, 1981,17(1). 185-203.
[26] Rekleitis I. Steerable filters and ceptral analysis for optical flow calculation from a single blurred image. Vision Interface,1996,12(1). 159,166.
[27]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. 2004,40(23). 15-19,49
[28]钱春强,王继成.基于改进约束最小二乘方法的图像复原算法.计算机技术与发展. 2007,17(6). 9-11,14
[29]李红阳.运动图象恢复.清华大学,2000. 18-31
[30]周玉,彭召意.运动模糊图像的维纳滤波复原研究.计算机工程与应用. 2009,45(19). 181-183
[31] M.M.Sondhi. Image Restoration: The Removal of Spatially Invariant Degradations. Proc. IEEE, 1972, 60(7).842-853.
[32]刘艳春.利用传播波方程作运动模糊图像质量恢复的客观评价.广州:暨南大学,2002.
[33]刘明.基于图像复原的航空摄影前向像移检测及补偿技术研究.中国科学院研究生院(长春光学精密机械与物理研究所),2005
[34]杨琬,吴乐华,范晔.数字图像客观质量评价方法研究.通信技术. 2008,7(41). 244-246
[35]倪晓明.基于小波变换图像质量评价新算法.厦门:厦门大学,2009.
[36]路文,高新波,王体胜.一种基于小波分析的部分参考型图像质量评价方法.电子与信息学报. 2009,31(2). 335-338
[37]安雪晶,田媛.无参考图像质量评价方法的设计原则.中国光学与应用光学. 2009,2(2). 140-144
[38] WANG X, TIAN B F, L IANG CH. Blind image quality assessment for measuring image blur. IEEE, 2008 Cong. Image S ig. Proces , Hainan, China, 2008. 467-470.
[39]蔡利栋.运动模糊图像恢复质量的客观评价:传播波方法.北京:第二届全国智能视觉监控学术会议,2003.
[40] Cai Lidong. Traveling Wave Equations and Restoration of Motion-Blurred Images. Acta Automatica Sinica,2003,29(3). 466 - 471.
[41]刘艳春,蔡利栋.基于传播波方程的图像恢复质量客观评价方法的验证.暨南大学学报(自然科学版).2004,25(1). 30-35
[42] Li-Dong Cai.Objective Assessment to Restoration of Global Motion-blurred Images Using Traveling Wave Equations. Multi-Agent Security and Survivability, 2004 IEEE First Symposium on.2004
[43] Cai L D,Zhen J W. Image blur caused by uniform linear motion and its removal using traveling wave equation and Hough transform. PROC of Second INT. CONF. on Image and Graphics,Wei Sui(Ed.),Hefei, China,SPIE, August 2002,268-275