基于深度学习的羽毛球正手高远球动作标准智能评估方法OA

Propose a deep learning based intelligent evaluation method for high and far ball movements of badmin-ton forehand(DL-BEFBFS).Input image data samples into an improved convolutional neural network,extract spa-tiotemporal features of badminton movements,and obtain intelligent recognition results of badminton forehand move-ments;Using the spatiotemporal characteristics of actions to calculate intelligent evaluation indicators such as frame height matching degree and center of gravity stability index;Assign weights to indicators based on their contribu-tion,and use principal component analysis to achieve intelligent evaluation of badminton forehand high and far ball action standards.The experimental results show that the action level evaluation accuracy of the DL-BEFBFS method is 0.97,which is 0.075 higher than traditional evaluation methods and has better evaluation performance.