健美操运动员高难度视频动作识别方法研究  

Research on Recognition Method of Aerobics Athletes’ Difficult Video Movements

作　　者：贺莉 李慧萌 金庆凯 赵树桐 HE Li;LI Huimeng;JIN Qingkai;ZHAO Shutong(School of Physical Education,West Anhui University,Lu ran 237012,China)

机构地区：[1]皖西学院体育学院,安徽六安237012

出　　处：《安阳工学院学报》2022年第6期121-125,共5页Journal of Anyang Institute of Technology

基　　金：安徽省体育社会科学研究项目“和谐社会构建中的城市体育发展”(ASS2016203);皖西学院校级人文重点“安徽省高校健美操运动队建设研究”(WXSK202126)。

摘　　要：健美操运动员动作视频识别中,由于高难度动作特征的表征效果较弱,因而识别精度下降。文章提出健美操运动员高难度视频动作识别方法,即以视频序列中连续两帧的图像作为目标,计算像素点对应的灰度差值,将其与设置的判断阈值相对比,区分前景与背景,提高动作特征的表征效果,检测出目标动作,提取出目标动作的特征,并对目标动作特征进行分类,完成健美操运动员高难度视频动作的分类识别。实验结果表明,计算复杂度低、识别精度高,实用性得到了进一步提升。In the process of action recognition of Aerobics athletes, due to the weak representation effect of difficult action features and the decline of recognition accuracy, this paper puts forward the research on high difficulty video action recognition method of Aerobics athletes. Taking two consecutive frames of video sequence as the target, the gray difference corresponding to the pixel is calculated and compared with the set judgment threshold to distinguish the foreground from the background, so as to improve the representation effect of action features. The target action is detected, the feature of the target action is extracted, and the feature of the target action is classified to complete the classification and recognition of the Aerobics Athletes’ difficult video action. The experimental results show that the computational complexity is low, the recognition accuracy is high, and the practicability is further improved.

关 键 词：健美操运动员 高难度视频动作 动作识别 特征提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]