中文体育类核心期刊

中国人文社会科学期刊AMI综合评价(A刊)核心期刊

《中文社会科学引文索引》(CSSCI)来源期刊

美国《剑桥科学文摘》(CSA)收录期刊

中国高校百佳科技期刊

基于KCF算法的足球运动员体能录像跟踪技术研究

The Video Analysis of the Players' Physical Fitness during the Soccer Competition Based on the Kernelized Correlation Filter

  • 摘要: 基于对现有视频跟踪技术的分析,提出一种基于KCF算法的多视角、多目标足球跟踪技术,通过调整学习率和样本范围提高准确率,并结合多视角视频信息加权定位球员位置。测试结果显示:该方法能够提高抗遮挡能力,保证跟踪处理速度,减少相似跟踪目标带来的误差,解决足球比赛中录像跟踪的多目标跟踪问题。

     

    Abstract: Through the study of the current video tracking technology, a new algorithm, named Kernelized Correlation Filter with multi-dimensional video analysis, can increase the accuracy by adjusting the study rate and samples' range, and track the players' movement combining with the weighted video information.The results show that the method can improve the blocking-resistance ability, guarantee the tracking processing speed, and reduce the tracking error of the similar target, which can better solve the soccer video tracking problem of more targets tracking.

     

/

返回文章
返回