Algorithm 1: Overview of the proposed video deblurring method.

Input: The blurry video. Divide the video into M groups that have N frames in a group. Set the group ordinal of the video m = 1 and the frame ordinal of this group n = 2. Repeat  Repeat (1)Obtain the first deblurred frame L1 of this group by utilizing an image deblurring method. (2)Perform motion estimation algorithm to get the motion vector between the blurry frames Bn−1 and Bn, and using it to derive the motion-compensated frame In from the previous deblurred frame Ln−1. (3)Obtain the preprocessing motion-compensated frame IP by preprocessing In, and then estimate the blur kernel k with IP and Bn by the regularization method. (4)Estimate the deblurred frame Ln by the spatiotemporal constraint algorithm with k and In. (5)n ← n + 1.  Until n > N m ← m + 1 Until m > M Output: The deblurred video.