US 12,257,478 B1
Method for assisting golf training and computer program product
Canrong Yao, Fujian (CN); Renyu Huang, Fujian (CN); and Yinzhong Chen, Fujian (CN)
Assigned to Xiamen Xingtel Artificial Intelligence Co., Ltd., Fujian (CN)
Filed by Xiamen Xingtel Artificial Intelligence Co., Ltd., Fujian (CN)
Filed on Sep. 4, 2024, as Appl. No. 18/824,213.
Application 18/824,213 is a continuation of application No. 18/788,219, filed on Jul. 30, 2024, abandoned.
Claims priority of application No. 202410864572.8 (CN), filed on Jun. 28, 2024.
Int. Cl. A63B 24/00 (2006.01); A63B 102/32 (2015.01)
CPC A63B 24/0006 (2013.01) [A63B 24/0021 (2013.01); A63B 2024/0015 (2013.01); A63B 2024/0056 (2013.01); A63B 2102/32 (2015.10); A63B 2220/05 (2013.01); A63B 2220/806 (2013.01)] 12 Claims
OG exemplary drawing
 
1. A method for assisting golf training, comprising:
capturing a video of a current swing by a user and corresponding simulator data, wherein the simulator data comprises golf hitting data and golf flight data of the current swing, the video is captured by a video acquisition device, and the simulator data is captured by a golf simulator;
analyzing a swing action in the video by using a plurality of preset human body key points related to the swing action to obtain an image frame sequence corresponding to each swing stage in the current swing and further obtain a human body key point set sequence corresponding to the image frame sequence, wherein each frame image of the image frame sequence is corresponding to a human body key point set of the human body key point set sequence, and a single swing action comprises a plurality of predetermined swing stages;
according to a relationship between the human body key points in the human body key point set sequence corresponding to each swing stage, determining a plurality of action features corresponding to a current swing action;
aligning the plurality of action features of the current swing with a plurality of action features corresponding to an optimal personal swing model preset for the user in time and space;
comparing one or more of the plurality of action features of the current swing with corresponding one or more of the plurality of action features corresponding to the optimal personal swing model to obtain an action difference between the current swing and the optimal personal swing model, wherein the optimal personal swing model is preselected from historical swing actions of the user or is a swing action of the user that is set by the user under guidance of a coach;
comparing at least one of the golf hitting data or the golf flight data of the current swing with at least one of golf hitting data or golf flight data corresponding to the optimal personal swing model, to obtain a swing result difference between the current swing and the optimal personal swing model;
according to a plurality of action features of a single swing and golf hitting data and golf flight data corresponding to the single swing, modeling the single swing to obtain a single swing model, and storing at least one of video pictures of the single swing, the plurality of action features of the single swing or the golf hitting data and golf flight data corresponding to the single swing as data comprised in records of the single swing model, wherein the single swing comprises the current swing; and
updating the optimal personal swing model according to at least one of training situations or assessment situations of the user.