Abstract

The game of golf is a popular sport not only for elite players, but also for a large number of recreational athletes. As in other sporting activities, improvements in performance are an important driver for playing. However, performance evaluation is perhaps the least investigated area of computer science within sports. Today, coaches make use of radar equipment, video and other tools, not only for detection of what to improve, but also for explaining what the problems are. Video is often restricted to 2D analysis and although there are sensor systems that can be attached to players and which can be used for 3D analysis, these systems are not, in general, easily accessible for recreational athletes. It is therefore of interest to investigate other approaches for 3D analysis which are more accessible to a larger segment of players, such as e. g. the Microsoft Kinect sensor. In the present work, the authors focus on two relevant questions related to modeling the golf swing using the Kinect sensor: First, is the resolution good enough (temporally and spatially) for capturing the golf swing and, second, how can we visualize the important parameters of the swing? In this paper the authors focus on the first of these questions and carry out initial investigations using the Microsoft Kinect sensor. Aus der Einleitung (geändert)