Identification of Piecewise Linear Uniform Motion Blur

Abstract

A motion blur identification scheme is proposed for nonlinear uniform motion blurs approximated by piecewise linear models which consist of more than one linear motion component. The proposed scheme includes three modules that are a motion direction estimator, a motion length estimator and a motion combination selector. In order to identify the motion directions, the proposed scheme is based on a trial restoration by using directional forward ramp motion blurs along different directions and an analysis of directional information via frequency domain by using a Radon transform. Autocorrelation functions of image derivatives along several directions are employed for estimation of the motion lengths. A proper motion combination is identified by analyzing local autocorrelation functions of non-flat component of trial restored results. Experimental examples of simulated and real world blurred images are given to demonstrate a promising performance of the proposed scheme.

Multiple-Image-Based Restoration for Motion Blur with Non-uniform Point Spread Function

“…Ranking the blurred images in the order of identification difficulty the single-image-based identification algorithm [4] for an initial identification of one PSF which seems to have the least difficulty to identify. A blur identification difficulty is proposed to rank an identification order of a provided set of blurred images, {g i } M−1 i=0 .…”

“…The PSF of the easiest image which tends to need the least information for identification should be firstly estimated. Then, other blur PSFs can be identified based on (4). Details of the ranking process and initial identification are presented in Sect.…”

“…As a result, the support number of PSF coefficients can be obtained. Section 2. linear model which have been introduced for the singleimage-based identification in our prior work [4]. By using the information from the motion direction and length analysis processes, the blur identification difficulty can be evaluated.…”

“…2. Firstly, to stimulate and strengthen strips along the blur direction, each blurred image is restored on trial by assuming the linear motion blur along a trial direction of φ and supposing that the distribution of PSF coefficients are forward ramp [3], [4]. Then, it is differentiated along the direction of 90 • + φ.…”

“…An original scene can be restored by using only one blurred image (single-image-based restoration) or by using additional images (multiple-image-based restoration). Most of the single-image-based approaches are limited for identification of simple blur models such as 1D motion models [1]- [3] or constant velocity motion models [1]- [4]. To identify more complex models of blurs, several methods such as auto-regressive (AR) model and maximum likelihood (ML) based methods [5], [6] have been proposed.…”