다중 카메라와 관성 센서를 이용한 골프 분석 시스템

Abstract

At present, analysis systems may be divided into two-dimensional (2D) and three-dimensional (3D) systems. A 2D analysis system utilizes the input images obtained by a camera and requires the application of image processing; 3D analysis system, on the other hand, utilizes a motion capture system and floor reaction force. A motion capture based analysis system requires at least 15 or more sensors on the user body, which should be the correct analysis heterogeneity feel to the user. In addition, the constraints of time and large space consumption add to the problems of this type of analysis method. The camera based method can produce significantly better results with only a few restrictions on the time and space, as compared to the motion capture system, it is difficult to analyze precise movements in a constrained environment. In this study, we propose a method for analyzing a golf swing using one inertial sensor, Kinect V2, and a high-speed camera. The inertial sensor is attached to the hand during the swing, and the fast movement is estimated using Kinect V2, which uses the SDK provided by Microsoft, with the user's body joint coordinates to extract the attitude data. In addition, the high-speed camera tracks and expresses golf shaft head.