摘要
基于对监控视频序列运动矢量分布的方向特性和起始点预测搜索技术的研究,提出了一种运动估计算法—自适应三分级搜索算法(ATLS),并结合Inter XScale微处理器架构对算法进行优化,解决了传统运动估计算法在嵌入式系统中搜索速度低的问题,算法运算复杂度低。实验结果显示,与运动矢量场自适应搜索算法相比,在图像质量几乎不下降的情况下,该算法平均搜索速度明显提高,综合性能优于运动矢量场自适应搜索算法(MVFAST)算法。
Based on the research about distribution of motion vector and search of predictive start, adaptive three levels motion estimation search algorithm (ATLS) was proposed, and was optimized with Inter XScale platform, which has solved motion estimation in low speed of conventional motion estimation algorithm. Simulation and experimental results indicate that the ATLS can increase the search speed while maintains the video performance in contrast to motion vector field adaptive search technique (MVFAST).
出处
《机电工程》
CAS
2008年第4期31-33,共3页
Journal of Mechanical & Electrical Engineering
关键词
INTER
XSCALE
视频监控
运动估计
块匹配
自适应三分级搜索算法
运动矢量场自适应搜索算法
Inter XScale
video surveillance
motion estimation
block matching
adaptive three levels motion estimation search algorithm (ATLS)
motion vector field adaptive search technique (MVFAST)