Local motion blur detection by Wigner distribution function

The movement of targets often lead to a noticeable motion blur in the local part of an image. Traditional methods mainly based on separating the blur part and performing Fourier transform are often affected by window effect due to the deviation of edge segmentation in the case of complex background and edge occlusion. Rather than applying a more sophisticated edge extraction algorithm, this paper presents a simple method which can guarantee the accuracy of PSF extraction and reduce the precision requirements in edge segmentation. The proposed method is based on Wigner distribution of image and can reduce the impact of local sampling and processing. With this method, the requirement of sampling interval can be reduced from the whole targets to the size of motion pixels. Besides, the application of Wigner distribution in Light Field can make it possible in the application of computational optical imaging.