中文摘要：

为了解决图像采集过程中因抖动引起的模糊问题,并能消除振铃效应,可对自相关性强烈的图像进行清晰复原,设计了模糊图像抖动轨迹估算函数;引入结构相似度,推导出降噪函数;提出了抖动轨迹耦合交替迭代最小技术的图像模糊消除算法.利用Bregman迭代法与交替最小技术,联合结构相似度,得到离散优化模型,对图像去噪处理;随后嵌入Wallis算子消除扩散效应,得到锐化图像;再由模糊图像与锐化图像,基于权重理论,估算模糊图像抖动轨迹;借助傅里叶逆变换,提取抖动轨迹的空间域;利用维纳滤波对抖动轨迹空间域规范化操作,获取清晰图像.测试了算法的性能.结果表明：与其他技术相比,算法消除自相关性强烈图像的抖动模糊效果更优异,可获得细节清晰的图像.

英文摘要：

In order to solve the problem of fuzzy induced by the jittering during the image acquisition,and effectively eliminate the ringing effect,as well as restoration the image of strong autocorrelation,the jitter trajectory estimation function of fuzzy image was designed; introducing the structure similarity and Bregman iterative method to deduce the Noise reduction function for proposing the image fuzzy elimination algorithm based on the jitter trajectory coupled the alternative iteration minimization technology.Used the Bregman iterative method and alternative minimization technology to combination with the structure similarity to get the discrete optimal model to eliminate the image noise,and then embedded Wallis operator to eliminate the diffusion effect for getting the sharpening image; according to the fuzzy image and sharpening image,the jitter trajectory of fuzzy image was estimated based on weight theory; then the spatial domain of jitter trajectory was extracted by the inverse Fourier transform; finally,the standardization was used to spatial domain by the wiener filtering to obtain the clear image.The performance of this algorithm was tested,and the results was showed that the blur eliminate effect of this algorithm was excellent,and it can obtained the image of clear details.