摘要
针对传统乒乓球动作识别未考虑动作的时序性和动态特性的问题,研究在二维DWT算法的基础上引入奇异值分解,设计出一种改进DWT算法,以获得乒乓球动作的局部特征,从而提升其识别准确率。实验结果显示:在识别准确率计算中,改进DWT算法的平均识别准确率为97.8%,优于其他算法,表明其识别精度较高;在识别时间计算中,改进DWT算法的识别时间与其他算法相比,均下降了50%左右,表明其识别效率较高。以上结果证明了改进DWT算法的优越性能,为乒乓球的动作识别提供了有效的技术支持。
With the advancement of modern technology,people’s research and analysis of sports have become increasingly in-depth.In response to the problem of traditional table tennis motion recognition not considering the temporal and dynamic characteristics of the motion,this study introduces singular value decomposition on the basis of two-dimensional DWT algorithm,and designs an improved DWT algorithm to obtain local features of table tennis motion,thereby improving its recognition accuracy.The experimental results show that in the calculation of recognition accuracy,the average recognition accuracy of the improved DWT algorithm is 97.8%,which is superior to other algorithms,indicating its high recognition accuracy.In the calculation of recognition time,the improved DWT algorithm has reduced the recognition time by about 50%compared to other algorithms,indicating its high recognition efficiency.The above results demonstrate the superior performance of the improved DWT algorithm and provide effective technical support for motion recognition in table tennis.
作者
赵莹
ZHAO Ying(Shenyang Conservatory of Music,Liaoning Shenyang 110818)
出处
《河北北方学院学报（自然科学版）》
2024年第5期34-39,44,共7页
Journal of Hebei North University:Natural Science Edition
关键词
乒乓球
二维DWT算法
奇异值分解
动作识别
Table tennis
2D DWT algorithm
Singular value decomposition
Action recognition
