Spatially-varying Motion Blur Detection And Parameter Estimation

Spatially-varying blur is caused by the relative motions between the camera and the objects of different motion speeds during camera imaging.Most of the existing algorithms are directed to global uniform blur,which is not consistent with real problems.This paper focuses on the problem of spatially-varying motion blur detection and parameter estimation.The main works and research results of this paper can be summarized as follows:1.Based on comparing the advantages and disadvantages of typical blur features,we improve the blur feature proposed by Su to make the scatter between classes bigger and the scatter within classes smaller.2.We propose a multiscale adaptive window blur region extraction algorithm,in order to solve the problem of support domain selection in local blur feature calculation and blur region segmentation threshold adaptive selection.We combine image segmentation results(such as color segmentation)with motion blur detection algorithms to improve accuracy.3.We combine the proposed multiscale adaptive window blur region extraction algorithm with the existing deblur algorithm,exploit our algorithm to realize automatic extraction of initial blur region and initial blur parameter estimation,then the latter algorithm is used to optimize the extraction and parameter estimation results of the spatially-varying blur,at the same time,image deblurring is achieved.The experimental results of simulation data and real data verify the effectiveness and robustness of the proposed algorithm.