期刊文献+

基于骨骼关键点检测的士兵训练动作分类研究 被引量:3

Soldier physical training movement classification based on skeletal key point detection
下载PDF
导出
摘要 人体骨骼关键点检测作为计算机视觉中的研究热点,产生了很多与之相关的运动辅助训练系统,并用于士兵体能训练的智能化。作为士兵体能辅助训练系统设计的基础工作,提出一种基于人体骨骼关键点的士兵体能训练行为分类方法。首先利用OpenPose模型对士兵体能训练运动视频进行骨骼关键点检测,然后针对体能训练运动的特点人工设计特征信息,最后利用改进的多分类SVM进行体能训练行为分类。实验结果表明,基于人体骨骼关键点检测的士兵体能训练动作分类方法能够快速有效地实现对视频的分类,准确率达到了92.9%。 By using human skeletal keypoint detection,a research hotspot in computer vision,many sports-assisted training systems have been generated,and applied to the intelligence of soldier physical training.A method of classifying soldiers′physical training behaviors based on human skeletal key points is proposed as a basic work for the design of soldiers′physical fitness-assisted training systems.Firstly,the OpenPose model is used to detect the skeletal key points of the soldier′s physical training movement video,then the feature information is designed manually for the characteristics of physical training movement,and finally the improved multi-classification SVM is used to classify the physical training behavior.The experimental results show that the soldier physical training movement classification method based on human skeletal key point detection can quickly and effectively achieve the classification of videos with an accuracy of 92.9%.
作者 曾文献 马月 李伟光 ZENG Wen-xian;MA Yue;LI Wei-guang(School of Information Technology, Hebei University of Economics and Business, Shijiazhuang Hebei 050000, China)
出处 《河北省科学院学报》 CAS 2022年第1期7-14,共8页 Journal of The Hebei Academy of Sciences
基金 河北省省级科技计划项目(20477601D)。
关键词 视频动分类 关键点检测 支持向量机 Video motion classification Key point detection Support vector machine
  • 相关文献

参考文献5

二级参考文献21

共引文献96

同被引文献36

引证文献3

二级引证文献3

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部