Golf practice video summarization using real swing detection

Abstract

Video camera is often used in many sports especially in golf to help athletes to improve their performance. Golfers often use video cameras to record their golf swing during golf practice session for golf swing analysis. In golf swing analysis, most golfers only focus on real swing because only real swings can truly show the problems of golfer’s swing. However, during golf practice session golfers may not record only real golf swing. Therefore, to shorten the process of golf swing analysis, golf practice video summarization method is developed in order to help golfers locate each real golf swing in golf practice video during golf swing analysis. This thesis proposed a method to summarize golf practice video recorded from face-on camera angle. Real swing can be detected by detecting takeaway and ball hit action of the golfer during golf practice session. The purposed method consists of five main processes which are golf club head detection, golf ball detection, takeaway detection, takeaway validation and ball hit detection. The result of experiments on golf practice video recorded with different cameras from face-on camera angle within different light conditions shows that this method can efficiently extract real swings. The precision and recall rate of the system are 90.78% and 92% respectively.