Image Restoration Based on Combined Dark Channel and Image Registration

Abstract

Due to the influence of water surface fluctuation and water turbulence, underwater target imaging detection technology faces great difficulties. This paper proposes an image restoration method combining dark channel prior theory and image registration technology, mainly using B-spline based non-rigid registration framework to correct image distortion. Since the conventional image registration algorithm often uses the average image of the sequence as a reference image, the average image tends to have severe blurring, which reduces the accuracy of registration. Therefore, this paper proposes to use the dark channel prior theory for the deblurring of the average image, and then use the deblurred average image as a reference for image registration. The test results show that the method can effectively improve the imaging distortion caused by surface turbulence.