摘要
当前在使用目标跟踪算法跟踪嵌入式篮球运动视频时,由于误差累积导致跟踪漂移且无法处理目标遮挡,造成目标跟踪算法执行时间较长和成功跟踪标记目标数量较少等问题。为此,提出了基于级联特征的嵌入式篮球运动视频目标跟踪算法。实验结果表明,目标跟踪算法跟踪执行时间较短,具有较高的时效性且成功跟踪标记目标的数量较多,能够准确跟踪篮球运动视频的标记目标,跟踪效果较好。
In view of the problems of the long tracking-learning-detection and the small number of successfully tracked targets caused by the error accumulation and the target occlusion,the paper puts forward the tracking-learning-detection of embedded basketball video based on Cascaded features.The result of the experiment shows that the tracking-learning-detection,which can ensure a short tracking time and a large number of tracked targets,has a good effect of accurately tracking the marked targets in the basketball video.
作者
黄国权
HUANG Guo-quan(Department of Physical Education,Wuhu Institute of Technology,Wuhu,Anhui 241003,China)
出处
《河北北方学院学报（社会科学版）》
2021年第3期51-55,共5页
Journal of Hebei North University：Social Science Edition
基金
安徽省教育厅高校人文重点研究项目(Wzyrwzd202006)。
关键词
级联特征
嵌入式
篮球运动视频
目标跟踪
cascade features
being embedded
basketball sports video
target tracking
