WOW: Wild-Open Warning for Broadcast Basketball Video Based on Player Trajectory

籃球運動中，空檔(Wild-open)的意思是有進攻球員未被成功防守。因此，這些發生空檔情況的影片通常代表進攻方戰術實行成功。在此論文中，我們設計了空檔偵測系統幫助教練以及球員得知自己的缺點以及比賽時可以施行的戰術，透過這些資訊，以達到賽前的情資蒐集以及賽後缺點檢討。經由可靠的相機參數偵測，球場偵測以及球員偵測與追蹤，於影片座標的球員位置得以轉至球場模型座標。我們定義三種空檔準則，當進攻球員滿足準則之一，系統將在畫面中標示出空檔球員位置，使觀眾能提高注意力在空檔球員身上。In basketball games, wild-open means that there is an offensive player not well defensed by his/her opponents. The occurrence of wild-open usually implies the existence of a successful offense tactic. In this paper, a Wild-Open Warning (WOW) system is designed to assist basketball coaches/players in revealing possible tactics of their opponents through watching the broadcast game videos. The system automatically extracts the semantic objects such as the court and the players in the video and calibrates players’ positions to match the real-world court coordinates. A robust and efficient algorithm for court detection and camera calibration is proposed for analyzing basketball videos. Wild-open event is detected when the position of an offensive player satisfies three predefined criteria. In the mean time, the system will mark the wild-open players to warn the viewers that they should keep high attention to certain players.Chapter 1 Introduction 1hapter 2 System Overview 5hapter 3 Video Pre-processing 7.1 Shot Segmentation 7.2 Noise Elimination 8.2.1 Court Mask Generation 10.2.2 Player Extraction 11.3 Top Boundaries Detection 12hapter 4 Court Detection and Camera Calibration 15.1 Court Side Determination 16.2 White court-line pixel detection 17.3 Line parameter estimation 17.3.1 Exhausted mode 18.3.2 Predicted mode 19.3.3 Generate configurations with quick rejection rules 20.4 Court model fitting 24hapter 5 Player Detection and Tracking 31.1 Player Candidate Extraction 31.2 Player Tracking 32.2.1 CamShift Tracking Algorithm 32.2.2 The Proposed Player Tracking Algorithm 34hapter 6 Wild-Open Warning Detection 37hapter 7 Experiment Results 39.1 Results for Calibration Algorithm 39.2 Results for Tracking Algorithm 42.3 Results for Wild-Open Warning 43hapter 8 Conclusion and Future Work 49ibliography 5