Real-time video stabilization algorithm based on efficient block matching for UAVs

In this paper, we proposed a video stabilization algorithm based on an efficient block matching on the airplane using Kalman Filtering. This algorithm uses the bit-planes to estimate and compensate the translational motion; while to compensate the rotation motion vector experienced in the video sequences we use the four local estimation approach to compute the rotational resultant vector. The global motion vectors of image frames are accumulated to obtain global displacement vectors, then they are filtered using the Kalman theory to get a vision stabilized system.