含噪运动模糊图像恢复  

Motion Deblurring Technology for Noisy Image

作　　者：赵志刚[1] 程姝[1] 吕慧显[1] 陈莹莹[1] 欧阳佩佩 潘振宽[1] 

机构地区：[1]青岛大学信息工程学院,青岛266071

出　　处：《青岛大学学报（自然科学版）》2014年第1期77-82,共6页Journal of Qingdao University(Natural Science Edition)

基　　金：山东省科学技术发展计划项目(批准号:2012YD01058)资助;国家自然科学基金(批准号:61170106)资助

摘　　要：提出一种新算法来估计含噪图像线性运动模糊的方向和长度。结合空间域及频域分析,使用Radon变换检测运动方向,双谱模型估计运动长度,最后利用Richardson-Lucy滤波恢复清晰图像。仿真实验结果表明,对于含噪图像该算法比以往算法更优胜。基本满足图像复原的稳定性、可靠性、精度高等要求。Motion blur is one of the most common blurs that graded images, an image restoration algorithm is established. degrades the images. In order to rest Restoration of such images is highly ore de- depen- ding on the blur parameters estimation. In this paper, we have proposed a novel algorithm to estimate mo- tion blur parameters such as direction and length. We used Radon transform to find blur direction and bis- pectrum modeling to find the motion length. Our algorithm is based on the combination of spatial and fre- quency domain analysis. Finally, using the Richardson-Lucy filter recover a sharp image. Experimental re- sults indicate that algorithm for noisy image is superior than ever more algorithm. It can satisfy the re- quirements of the image restoration which is stability, reliability and high precision.

关 键 词：运动模糊 RADON变换 双谱模型 含噪图像 参数辨识 复原滤波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]