基于Hough变换的运动模糊图像参数估计与恢复

The estimation and discovery of motion-blurred image based on Hough transform

在日常生活中,人们用相机拍摄景物时,经常会出现相机与景物之间存在相对运动的现象,这就使所拍摄图像产生了运动模糊。对于运动模糊的图像处理,一般是先找到图像的退化模型,然后对图像采取相反过程进行复原。而这其中的关键在于退化模型中运动方向和运动长度两个参数的估计。文中针对这一问题,采用基于Hough的运动模糊图像参数估计方法对运动方向进行估计,进而通过Radon变换得到运动长度,最后通过维纳滤波进行图像复原。结果表明该方法在一定程度上可以进行运动模糊图像的有效恢复。

In our daily life,motion-blurred image is caused by the relative motion between a camera and the object when shooting objects. For processing of motion-blurred,the usual way is to find the degradation model of an image,and then to use the inverse process to discover the image. The key of the processing way is to estimate two parameters which are motion-direction and the motion-scales. Aiming at the problem,this article uses Hough transform to estimate the motion-direction and then uses Radon transform to estimate the motion-scales. Finally,the discovery is processed by the Wiener filter algorithm. The result shows that the method can realize the effective discovery of motion-blurred image in a degree.