基于倒频谱的运动模糊图像PSF参数估计  

Estimating Blur Parameters of Point Spread Function of Motion-Blurred Image Based on Cepstrum

作　　者：谢伟[1] 秦前清[1,2] 

机构地区：[1]武汉大学国家多媒体软件工程技术研究中心,武汉市430079 [2]武汉大学测绘遥感信息工程国家重点实验室,武汉市430079

出　　处：《武汉大学学报（信息科学版）》2008年第2期128-131,共4页Geomatics and Information Science of Wuhan University

基　　金：国家973计划资助项目(2006CB701303);国家863计划资助项目(2006AA12Z132)

摘　　要：针对任意方向的直线运动模糊图像，说明只能直接在运动方向上得到运动参数，从而设置二维点扩展函数（point spread function，PSF）。在此基础上，应用倒频谱分析法给出了PSF参数估计的方法。实验表明，该方法在模糊为任意方向且模糊范围介于5～55像素时对参数的估计误差较小，能保证较好的恢复质量；当模糊范围超出该范围时，估计误差急剧加大，估计值不可信，无法保证恢复质量。The method of establishing the point spread function （PSF） of motion-blurred image is discussed. For linear motion blur, it is proved that the two dimensional PSF must be established directly in the motion direction, instead of decomposing to two axes or rotate to the horizontal axis. Then an algorithm is presented to estimate the two key parameters of PSF： establish the cepstrum of the motion-blurred image and find the minimum of it, the location of the minimum gives the information of blur range and blur direction. Experimental results show that the method can guarantee a small error a good quality of restoration when blur is at arbitrary direction and between 5-55 pixels. A big error occurs when blur range is out of 5-55 pixels.

关 键 词：图像恢复 运动模糊 点扩展函数 倒频谱 

分 类 号：P237.3[天文地球—摄影测量与遥感]