摘要
由相机抖动引起的图象退化是一种常见现象,由于退化函数和隐藏图象均未知所以对单幅模糊图象进行恢复具有很大难度。介绍了一种新的模糊核估计方法和贝叶斯最小均方误差采样反卷积算法,提出了一种同时具有该模糊核估计和贝叶斯最小均方误差采样算法功能的单幅图象去运动模糊方法,并使用由相机抖动引起的运动模糊图象进行了实验验证和分析,实验结果表明,该方法恢复结果中保留了图象重要边缘信息,振铃失真很少,视觉效果较好。
Camera shake is a common source of degradation in photographs. Restoring blurred pictures is challenging because both the blur kernel and the sharp image are unknown. A algorithm based on a new method of estimating blur kernel and Bayesian minimum mean squared error Sampling deconvolution algorithm is introduced. Then, an image restoration method of motion-deblurring is presented by combining them, and is used to process some motion-blurred photographs caused camera shake. Experimental results show that this method can preserve important edges of images, little ringing artifacts, and the visual quality of restoring images is better.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第11期4305-4308,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60975016
61002052)
浙江大学CAD&CG国家重点实验室开放课题基金项目(A1214)
海军大连舰艇学院科研发展基金项目(20110006)
关键词
去运动模糊
贝叶斯最小均方误差
模糊核估计
反卷积
图像增强
motion-deblurring
bayesian minimum mean squared error sampling
kernel estimation
deconvolution
image en-hancement
