Digital image translational and rotational motion stabilization using optical flow technique

Abstract

This paper proposes a new digital image stabilization (DIS) systems based on optical flow technique. Unlike previous DIS systems developed mainly for removing the translational motion disturbance, the proposed system removes not only the translational but also the rotational motion disturbances. A computational scheme that facilitates the local motion vector field to estimate the global disturbing translational and rotational motions is developed. First, the optical flow technique is used to estimate the local motion vector field of the image, yielding the velocity of each pixel in the current image frame. Then, the global translational and rotational motion parameters are determined in terms of the least squares estimation. Finally, these motion vectors are used to generate the counterbalance signals for removing the disturbance motion. Owing to the additional ability of the rotational motion removal, the new DIS system suppresses the undesirable translational and rotational disturbances effectively and thus enhances the DIS performance significantly.