摘要
提出一种基于像素特征的运动块补偿算法.该算法充分利用视频图像的亮度特征与运动关系,实现了基于目标亮度特征进行分割运动块.同时针对视频图像目标运动特点,求得相应的运动方程,更加准确的反映了目标物体的真实运动,得到较为真实的运动矢量场.实验结果表明,与传统的快速分类搜索(CFS)、全搜索法(FS)相比,使用本文算法得到的视频图像帧,具有更加真实的运动矢量场,并且PSNR值相对CFS大约有20 dB的提高.
A motion block compensation algorithm based on characters of pixel is proposed in this paper. A motion region-segmentation which adequately utilizes the relationship with luminance characters and motion properties in the video frames is presented. Meanwhile, according to the characters of motion object in the video frame, the equations of motion are obtained for showing the real motion field. The experimental results show that the image that is estimated by novel algorithm not only has more genuine motion field, but also PSNR has been improved about 20 dB compared with CFS.
出处
《北京交通大学学报》
EI
CAS
CSCD
北大核心
2008年第2期58-62,共5页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
国家自然科学基金资助项目(6077204060772042)
关键词
亮度特征
运动块
运动方程
运动补偿
luminance character
motion block
equation of motion
motion compensation
