期刊文献+

基于Kinect2.0的健身动作训练系统 被引量:1

Fitness exercise training system based on Kinect v2.0
下载PDF
导出
摘要 为减少在健身训练指导中存在的主观因素影响,促进健身运动的标准化、规范化,提出通过Kinect获取人体主要关节的运动,进一步转化为表征运动的线条图。利用Harris角点检测得到特征点,基于MOPS描述符进行特征匹配,采用等特征点比例分段法进行分析,计算各项指标,为学员提供有效地建议。实验结果表明,对于一系列连续动作,对比结果与人为对比相比在细节方面更加优秀。 In order to reduce the influence of subjective factors in fitness training guidance and promote the standardization and standardization of fitness exercises,it is proposed to use Kinect to obtain the movements of the main joints of the human body,and further convert them into line diagrams that represent sports. Feature points were obtained using Harris corner detection,feature matching was performed based on MOPS descriptors,and equal feature point proportion segmentation methods were used to analyze and calculate various indicators to provide effective suggestions for students. The experimental results show that for a series of continuous actions,the comparison results are better in detail than the artificial comparison.
作者 刘长喜 付少锋 LIU Changxi;FU Shaofeng(School of Computer Science and Technology,XiDian University,Xian 710126,China)
出处 《智能计算机与应用》 2020年第4期35-39,共5页 Intelligent Computer and Applications
基金 国家大学生创新创业训练计划(201910701098)
关键词 KINECT 连续动作对比 训练辅助 特征匹配 关节姿态 Kinect Continuous motion comparison training assistance feature matching joint pose
  • 相关文献

参考文献11

二级参考文献111

共引文献102

同被引文献5

引证文献1

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