图像模糊度参数估计与图像复原的实验及分析

The Experiments and Analysises for the Estimates of Blur Extent Parameter and the Restoration of the Blurred Image

由于成像系统与运动物体之间存在相对运动,摄像机在成象过程中经常会产生运动模糊。为了有效复原退化图像,需要从模糊图像中估计模糊函数的某些参数、噪声统计量等相关信息。在运动模糊中,模糊函数中最重要的参数是模糊度参数。模糊图像的模糊度内的像素点存在着相关性,利用其相关性可估计模糊度参数。根据运动模糊图像中沿着运动模糊方向与其他方向有不同图像特性的特点,估计水平直线运动图像模糊度参数并采用维纳滤波方法复原模糊图像,取得了较好的恢复效果。

Motion blur is be produced during taking photos, because of opposite motion between camera and object. In order to restore blur image effectively, certain parameter of blur function and statistical value of noise have to be estimated. The most important parameter of blur function is the blur extent parameter in motion blur. There is the correlation between pixels in the blur extent of blurred image, which is applied to estimate the blur extent parameter. In the experiment, we determine the blur function of linear uniform motion blurred image by using the estimated blur extent and restore the blurred image by using Wiener filter as well as have got the satisfying restoration results.