基于RPCA的局部运动模糊图像的复原

Restoration of Local Motion Blurred Image Based on RPCA

针对视频中局部运动模糊图像复原问题,首先运用RPCA将运动物体从图像中分离出来,用傅里叶变换得到其频谱图,对频谱图进行边缘检测和Hough Transform,估算出PSF值,然后用维纳滤波进行图像恢复并合成到背景中。从实验结果可以看出,该方法能够很好地分离图像的前景和背景,并将恢复后的前景图像融合到背景中,没有振铃效应产生,取得了很好的复原效果。

Aiming at the local video motion blurred image restoration problem,we use RPCA to separate moving objects from the image,the spectrum is obtained by Fourier transform,edge detection and Hough Transform on the spectrum,estimate the value of PSF,and then use the Wiener filter for image restoration and synthesis into the background.The experiment results show that this method can well separate the foreground and the background image,and the prospect of the restored image into the background,there is no ringing effect,achieved good restoration effect.