【摘要】 针对UMHexagonS算法的运算量大、耗时长、复杂度较高等问题,提出了改进方案,采用混合时空域起点预测,提高起始点的预测精度;为搜索模块设定阈值,提前终止搜索;在整个搜索过程中,提出了一种搜索模块象限区域分割法,有效地减少了搜索点数。实验结果表明,改进后的算法在保证信噪比和控制误码率的情况下,比原算法减少了14%~38%的运动估计时间。更多还原

【Abstract】 As UMHexagonS algorithm has the disadvantages of big computation,time consuming and high complexity,this paper proposes some schemes of improvement,which mainly includes three aspects.Firstly,use spatio-temporal initial point prediction to improve the forecast accuracy of the initial point;Secondly,set threshold for the search module to terminate search;Thirdly,put forward a search module quadrant regional segmentation method in the whole search process,effectively reducing the searching points.Through the experimental results show that the improved algorithm can in guarantee SNR and bit-error rates than the original algorithm,reduce 14%~38% of the motion estimation time.更多还原