以視覺為基礎之高爾夫球自我訓練系統

Abstract

完美且熟練的揮杆對於高爾夫球這項運動是非常重要的，大部分的球員花費大量時間用以精進各種揮杆技巧。然而，若沒有教練的監督或指示之下自行練習，球員容易因為不正確的姿勢造成沒有效率的揮杆方式，甚至造成運動傷害。本論文提出了以視覺分析為基礎的高爾夫球自我訓練方法，目的在於分析使用者的姿勢是否正確，並且能夠即時回傳給使用者。首先，我們提出兩個能夠即時影像處理的技術，輪廓提取和骨架計算，用來產生特徵線。接著，我們導入專業高爾夫球的知識，利用各種自動產生的輔助線來判斷目前姿勢的正確性，並且利用視覺化及聲音的矯正指示，以便使用者能夠更加迅速地調整姿勢。同時，也可以在揮杆以後，重新觀看自己動作。為了讓使用者方便，我們也在智慧眼鏡上面實作，先透過電腦作影像處理，然後把所偵測到的特徵線還有如何改正姿勢的資訊會被傳送到智慧眼鏡上面，讓使用者能夠立即調整自己的動作。

Since a proficient golf swing is a key element for golf, golfers always make effort to improve their swing skills. However, if a golfer practices himself/herself without the supervision of a coach, he/she cannot effectively perceive postural faults in swings and improve the swing mechanisms. Moreover, improper postures may even cause harm to muscles and ligaments of the body. In this thesis, we propose a vision-based golf training system, named “Improve My Swing,” which aims at improving the golfer’s swing skills. We present two computation-efficient image processing algorithms, contour extraction and skeleton computation, to automatically detect feature lines. Then, we evaluate whether the golfer’s body is properly aligned based on professional golf training knowledge and provide instructions of posture adjustment to the golfer in a more intuitive and unambiguous manner. Different audio alerts will be given to the golfer according to which posture fault is detected. Images of incorrect postures with superimposed and visualized instructions are also saved for the golfers to review their own swings. For the user’s convenience, we also implement our system on Google Glass. After running our system with image processing, feature lines and instructions of posture adjustment are sent to Google Glass. The users can adjust their postures through the visualized instructions on Google Glass.