运动估计是视频编码中最重要且最耗时的一部分,它占用整个视频编码60%~80%的时间。研究高效的、快速的运动估计算法是目前视频压缩技术中的重要研究课题。基于H.264视频编码标准,选择x264作为测试编码器,分析了x264的4种运动估计算法,通过加入非对称小菱形搜索,降低搜索点数,部分算法优化,对非对称十字型多层次六边形格点搜索算法(UMHexagonS)进行了改进,提高了运动估计算法效率。提出了非对称十字型多层次八边形格点搜索(x264_ME_UMO)算法。通过对各种视频序列的测试表明,在基本保持原有编码性能和图像质量的情况下,优化后的算法编码速度平均提高了约17%,能更好地满足实际应用的需求。
Motion estimation is the most important and time consuming part of H.264 video coding standard,it takes about 60% -80%of encoding time. Study on high performance, fast motion estimation in video compression technology now is an important research topic. Based on the H.264 video coding standard, select x264 as the test code, the paper has analyzed the x264 four kinds of movement estimate algorithm, through joins the asymmetrical small diamond search, reduces the search points, the partial algorithms optimize, has made the improvement to asymmetrical cruciform multi-level hexagon grid point search algorithm (UMHexagonS), enhances the movement estimate algorithm efficiency, proposes the asymmetrical cruciform multi-level octagon grid point search algorithm (x264 ME UMO). Testing with a variety of video sequences proves that, remained of the original coding performance and image quality and the encoding speed average increase of about 17%, to better meet the needs of practical application.