青少年体育评估的动态姿态分析方案设计

A Dynamic Posture Analysis Method for Youth Sports Evaluation

目前中学生体育测试采用的人工评估方案存在评估效率较低、易受到外界因素的干扰且无法对学生运行记录进行追溯的问题。为解决这些问题,本文提出一种智能体育评估方案。方案采用基于Alphapose的姿态估计算法对视频内容进行解析得到由人体骨骼关键点所组成的时间序列,并利用运动序列数据构建了基于动态时间规整的运动序列分类模型来判断学生的运动姿势是否标准。实验结果表明,本文的模型在多场景、多目标和多视角等情况下均有较高精度。

At present,the manual evaluation scheme used in physical education testing for secondary school students has problems such as low evaluation efficiency,susceptibility to external factors,and inability to trace student operation records.To address these issues,this article proposes an intelligent sports evaluation scheme.We used an Alphpose based pose estimation algorithm to analyze video content and obtained a time series composed of key points of human bones.We also constructed a motion sequence classification model based on dynamic time warping using motion sequence data to determine whether a student's motion pose is standard.The experimental results show that the model proposed in this paper has high accuracy in multiple scenarios,targets,and perspectives.