Analysis of multiple moving objects in video for removing motion blur

摘要

Recently, various digital image acquisition systems have been used to capture a scene. In those systems, image degradation, such as motion blur is unavoidable. Motion blur is caused by relative motion between objects and background. In this paper an image degradation model for multiple objects with different motions is proposed. The proposed model describes the nature of motion blur and the relationship with background. We also proposed a space-variant iterative image restoration method with an accordingly proper segmentation algorithm for removing the motion blur.