Abstract:  In order to reduce the computational complexity of block matching in video compression, and avoid local optimum of block matching, while maintaining a high compression ratio, a new fast block matching algorithm was proposed. First, determine whether the block is a static block. If it so, stop searching of this block. If not, do motion vector prediction to determine a searching center, from where the search begins and the route will be from the inside out. Meanwhile, a stop condition is combined to stop searching in an early time and an acceptable position. The proposed fast search algorithm for block matching was compared with full search and diamond fast search through simulation to prove its superiority. The above improved motion estimation algorithm was applied in a robot vision control system. The experimental results show that among full search, diamond fast search, and fast search algorithm proposed in this paper, the overall performance of the proposed algorithm is the best of the three methods.  

Key words: video compression, block matching, fast search, wavelet transform, remote vision control