An effective motion-blurred image restoration approach for automated optical inspection

Abstract

This paper presents a real-time restoration method for linear local motion-blurred images in automated optical inspection (AOI). The proposed approach is to firstly divide such an image into many sub-images and then detect the blurred sub-image by the gradient distribution and the cepstrum maximum. For a blurred sub-image, the blur direction and blur length are estimated in order to calculate the parameters of the point spread function (PSF). The Richardson–Lucy deconvolution algorithm and Wiener filtering are employed to restore this blurred sub-image. Through experimentation, the proposed algorithm produces good results on blurred images caused by the linear motion AOI equipment. To test its performance, the proposed algorithm is compared with other approaches by using a real captured printed circuit board (PCB) image, and it is proven to be superior to the others in terms of accuracy.