面向体育测试的人体运动姿势实时识别研究

Research on Real-time Human Motion Posture Recognition for Sports Testing

针对传统体育测试方式需要专业的测试设备和人员,存在人工成本较高和人工计数不准确的问题,研究一套智能体育测试系统以帮助高校更精准、更高效地得到体育测试结果。为了提高识别精度,提出一种基于OpenPose算法改进的人体运动空间融合姿态识别分析模型Pos-RSF,从而实现了对体育测试人体运动姿势估计的快速准确识别,提高了模型运行效率。实验表明,该模型具有更高的准确性和实时性,其总体得分PCKh@0.2为91.2%,与当前先进技术相比有显著提高,能够满足体育测试基本要求。

The traditional sports testing method needs professional testing equipment and personnel,which has the problems of high labor cost and inaccurate manual counting.Therefore,the purpose of this paper is to study a set of intelligent sports test system to help colleges and universities to get more accurate and more efficient sports test results.In order to improve the recognition accuracy,we proposed a model of pos⁃ture recognition of spatial fusion(Pos-RSF)based on the improved OpenPose algorithm,which could realize the fast and accurate recognition of human motion posture estimation in sports tests and improved the operating efficiency of the model.The experiment showed that the posture recognition model we studied has higher accuracy and better real-time,the overall score PCKh@0.2 was 91.2%,a significant improvement over the most advanced technology,and can meet the basic requirements of sports testing.