Design And Implementation Of Intelligent Management System Of Sports Video On The Mobile Terminal

In recent years,people begin to pay more attention to sports,and have higher and higher requirements for the quality of sports.Professional training institutions can detect and analyze various data in sports and help people improve the quality of sports.But this approach often costs more,which is not conducive to popularization.With the boom of mobile Internet,some sports fitness apps have gradually come into the public’s view.The functions of this kind of app are mainly teaching.Their disadvantage is the lack of targeted analysis of user movement,which may lead people to exercise in non-standard ways for a long time.The above problems will be solved by implementing a mobile terminal application that analyzes the motion process based on video.In addition,with the help of mobile terminal intelligence,recording sports video accurately by realizing motion recognition will minimize network flow and improve the accuracy of video analysis.This thesis investigates the development status of human pose estimation,motion recognition technology and iOS platform related technologies.Combined with the requirements of motion analysis and video management,it designs the system architecture,functional modules,data model and system interface.It studies and designs the pose matching algorithm for motion recognition,and finally develops and tests intelligent management system of sports videos on the mobile terminal.The main contents of this thesis are as follows:1.Requirement analysis and design of system.Based on the study background and study goals,this thesis subject analyzes the core business process of the system,and describes the requirements of front-end and expansion on the core business process.On this basis,we design the whole system,design the system function module according to the function,and design the function flow,implementation scheme,required data model and system interface in the module.2.Research and design pose matching algorithm.Aiming at the demand of sports recognition from camera video stream,we propose a pose matching algorithm based on support vector machine,which is used to judge whether the human posture in the image matches the target posture.Based on this algorithm,the automatic recording of motion recognition is realized in the system implementation stage combined with the timing relationship of key frames of motion video.3.Implementation and testing of iOS application system.According to the requirements of the system,this thesis explain the design scheme.The scheme includes the implementation process of the UI interface and business logic corresponding to each functional module.It also include the rational design of the page hierarchy according to the relationship between modules.The common parts of the modules are encapsulated.Finally,the thesis subject tests the function,performance and stability of the system application.