摘要
为解决常用的基于能量谱和倒谱零点的运动模糊图像复原方法的噪声敏感问题,提出了一种高鲁棒性和适应性的新算法。该算法使用中值互补模板法进行噪声能量的估计,在对数能量谱域内使用改进的Radon变换实现模糊角度的识别,通过旋转图像使运动模糊出现在水平方向,然后在双谱域内识别模糊长度,大大降低了噪声对识别的影响。实验结果表明:本文算法能够处理信噪比大于5 dB的噪声模糊图像,图像复原效果显著。
The conventional restoring method of motion blurred image, based on power spectrum or bispeetrum, has high sensitivity to noise. In order to overcome this shortcoming, this paper proposes a new method, in which a median complement template is adopted to estimate the noise power. After the motion direction is identified in logarithm power spectrum domain, the blurred image is rotated to horizontal direction. Moreover, the motion extent is calculated in bispectrum domain of the rotated image such that the effect of noise on the identifying precision can be greatly reduced. It is shown via experiment results that the proposed method can effectively identify the motion parameters from noisy blurred image with more than 5 dB signal-to-noise ratio, and attain a better restored image.
出处
《华东理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2012年第4期506-510,共5页
Journal of East China University of Science and Technology
关键词
图像处理
运动模糊
点扩散函数(PSF)
RADON变换
双谱
image processing
motion blur identification
point spread function （PSF）
Radon transformation
bispectrum
