Research On Digital Television System Backend Motion Estimation Algorithm

Based on the digital TV video processing technology,this paper presents the following improved algorithm based on the detailed analysis and comparison of the motion estimation algorithms existing in frame rate up-conversion.1)Based on the 3-dimension recursive search algorithm(3DRS),the luminance and Chroma signals(YC)are blended to improve the accuracy of the motion estimation vector.2)To increase the selection of dynamic candidate vectors and penalties,speed up the process of convergence,faster approaching the true direction.3)Increase the search window,can get more candidate blocks,and thus speed up the convergence rate.By dividing each frame into 6 × 2 macro-block,different search directions are taken for every macro-block,and the best candidate vector can be found more quickly.Its convergence speed and accuracy greatly accelerated.In addition to the forward motion estimation and the backward motion estimation in the conventional motion estimation algorithm,a bi-directional motion estimation algorithm is added to calculate the smooth motion vector field.The experimental results show YC hybrid large-search multi-phase motion estimation algorithm proposed in this paper is superior to both the objective evaluation and the subjective evaluation in terms of the accuracy of the motion vector estimation,the comparison of the algorithm execution time and the convergence speed.