基于AR的排球运动员击球姿态特征提取

Feature extraction of volleyball players’hitting posture based on AR

为提高排球运动员击球姿态修正能力,提出新的基于AR的排球运动员击球姿态特征提取方法。构建排球运动员击球姿态AR图像采集模型,提取边缘轮廓特征,分析排球运动员击球姿态的动作特征量,通过AR视景空间重构技术完成排球运动员击球姿态图像融合和特征分解。利用关联模糊度检测的方法,提取图像角点特征和灰度不变矩特征量,跟踪识别提取排球运动员击球姿态图像动作。实验结果表明,所提方法精度较高,动作跟踪稳定性较好,能够有效指导排球运动员击球姿态动作的修正和训练。

In order to improve the ability of volleyball players to correct their hitting posture,a new feature extraction method based on AR is proposed.The AR image acquisition model of volleyball players’hitting posture is constructed by extracting the edge contour features,and analyzing action characteristics of volleyball players’hitting posture.The image fusion and feature decomposition of volleyball players’hitting posture is completed by AR visual space reconstruction technology.The experimental results show that the proposed feature extraction method has high accuracy and good tracking stability,which can effectively guide the correction and training of volleyball players’hitting posture.