| The long-lasting prosperity of the Chinese table tennis team is an important factor in the ping-pong as a national ball inherited from China.In order to maintain the high level of competition of the Chinese table tennis team,the analysis of the game based on the game video and training video is an important means of technical and tactical analysis.With the continuous deepening of computer applications,the use of video and image processing technology to assist in the analysis of sports competition techniques and tactics has been widely recognized by the sports community,and several softwares have been developed for different sports projects to assist in technical and tactical analysis.The Software Engineering Laboratory of North China University of Technology has long been committed to the computer aided tactics analysis of the Chinese table tennis team,and has made due contributions to the national team's brilliant achievements.According to the actual needs of the national table tennis team and the status quo of video analysis technology,this paper presents two kinds of table tennis technical and tactical video analysis techniques,namely,"dot" semi-automatic analysis technology and video intelligent analysis technology.Among them,the "dot point"analysis technology is based on the pre-made analysis template,and uses the mouse click to obtain the video clip of the table tennis technique and tactics,thereby achieving the purpose of the athlete's technical motion recognition analysis,solving the problems of large workload and low recognition efficiency in the traditional manual recognition..Video intelligent analysis technology is to break down the video of table tennis competition into several key frame images,and then based on the traditional Bag Of View Word(abbreviated BOVW),combined with the characteristics of table tennis sports,through the use of SIFT features in the model.The traditional algorithm such as descriptor and K-means clustering combined with the background region feature point culling algorithm proposed in this paper completes the automatic recognition of athletes' technical movements.Through good design and implementation,this topic combines the "dot"semi-automatic identification technology and intelligent recognition technology to realize the "ping pong military division" human motion recognition system,which provides users with a simple interactive interface and also for human motion recognition personnel.Provides a convenient and easy-to-use management interface.Through the system test,the two identification methods are compared and tested.The experimental results show that the method proposed in this paper can identify the human body movements of table tennis players and achieve the expected indicators.The design and development of a similar human body motion recognition system for sports projects has certain reference value. |
PDF Full Text Request