融合运动幅度与方向特性的快速运动估计算法  

Fast motion estimation algorithm based on range of motion and directional characteristics

作　　者：万智萍[1] 

机构地区：[1]中山大学新华学院,广东广州510520

出　　处：《计算机工程与设计》2014年第2期588-592,635,共6页Computer Engineering and Design

摘　　要：针对目前运动估计算法搜索点过多与精确度低等现象,根据视频图像中运动块的运动幅度大小与其特有的方向特性,对运动估计算法进行改进。通过引入阈值的形式,快速而有效地对视频图像进行分类,并根据搜索点的分布情况,采用方向搜索模式对其进行分类搜索来减少所需搜索的点数。JM测试平台检验结果表明,该算法能够在保证视频质量的情况下实现以上目标,具有精度高、耗时少的市场优势。Because there are too many search points in the motion estimation algorithm and the accuracy of the algorithm is low. Motion estimation algorithm based on the range of motion and directional characteristics is improved. The threshold value on this mostion estimation algorithm is introduced to classify the video image. And according to the distribution of the search points, the direction of the search mode is used to do category search to reduce the number of points required by the search. Using the JM test platform to test, the test results show the algorithm can achieve high quality of the video with the market advantages of high precision and less time-consuming.

关 键 词：运动估计算法 视频压缩 阈值 运动幅度 方向特性 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]