一种改进的运动模糊图像恢复算法  

An Improved Motion-Blurred Image Restoration Algorithm

作　　者：李阳[1] 张路[1] 张志勇[1] 赵菲[1] 

机构地区：[1]国防科技大学ATR国防科技重点实验室,湖南长沙410073

出　　处：《计算机仿真》2011年第5期254-257,共4页Computer Simulation

基　　金：国家"863"重大专项(2008AA801106)

摘　　要：在运动图像复原的研究中,运动模糊图像的参数估计与恢复算法是近年来数字图像处理领域研究的热点问题。针对运动图像受到噪声信号的影响,为了提高图像恢复效率,通过对运动模糊图像的恢复过程进行分析,提出了一种改进的参数估计与图像恢复算法,即利用经阈值化处理的Radon变换估计模糊方法,通过微分自相关法估计模糊长度,最后应用带最优窗的维纳滤波进行图像复原。进行仿真的结果表明,改进算法能够较为精确地估算出运动模糊图像的模糊参数并取得了较好的恢复效果,提升了图像恢复的抗噪性能,具有实际参考价值。The estimation of parameters and the restoration of motion-blurred images are widely researched in the field of digital image prosessing at present.By analysising the prosess of restoration,an improved parameter estimation method and an advanced restoration algorithm are given.The direction is estimated by Radon transform algorithm with threshold.the length is estimated by differential-autocorrelation.and finally,the image is restored by wiener filtering algorithm with optimized-window.Experiments show that the integrated algorithm can accurately estimate the parameters of motion-blurred images and obtain better rescovery results.Moreover,the algorithm has better antinoise capability and can be used in reality.

关 键 词：运动模糊图像 芮登变换 自相关 图像恢复 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]