Design And Implementation Of Sports Video Recording And Storage Analysis System Based On Depth Information

With the rapid development of mobile smart terminals and people's attention to fitness,the market for sports analysis software is getting hotter.Most of the existing motion analysis software is limited by the venue and hardware equipment and cannot be widely used.Mobile smart terminals have the advantages of portability and high penetration rate,which are suitable for a variety of scenarios,and the hardware equipment configuration is becoming more and more abundant,such as multi-lens cameras and sensors.Based on these sensors,using technologies such as stereo vision and augmented reality,the collection of sports information that previously required special equipment can be transferred to mobile smart terminals,which is of great significance in practical applications.The motion video acquisition and storage analysis system based on depth information proposed by the author of this article takes the user's motion video as input,uses mobile phone related sensors to collect additional motion information,and uses wireless communication technology to upload the video to the server for analysis.Finally get the result videos from the server and bound them to the original videos so that they can be displayed on the mobile terminal.In this article,the author first analyzes the research background of the subject and clarifies the research objectives and content.Afterwards,the author proposed a method of-using binocular cameras,acceleration sensors and other equipment to collect motion information on the mobile terminal based on stereo vision,augmented reality and deep learning technology.The author introduces related technologies in combination with practical applications,and analyzes the requirements of the system,and determines the functions to be completed.According to the demand analysis,the author elaborated on the overall design and process of the system and module division.Subsequently,the author introduced the detailed design and implementation of each module one by one,and performed functional tests on each module to verify the validity and expound the scalability of the system.Finally,the author integrates the modules,completes the system and summarizes the development process,and looks forward to the future.The video capture and analysis system proposed in this thesis uses stereo vision and augmented reality technology to supplement the motion information lost when the 3D scene is projected to the 2D video in the traditional video shooting method.It solves the problem that it is difficult to analyze the user's actions from a single video information in the scene without identification equipment or infrared devices.According to the test results of the system,it is proved that the system can obtain additional motion information to assist the function of motion analysis.