基于条件随机场的体育比赛视频目标实时定位研究

Research on Real Time Target Location in Sports Video Based on Conditional Random Field

在体育比赛视频目标定位中,由于目标变化具有随机性,导致定位难度较大,为此提出基于条件随机场的体育比赛视频目标实时定位方法。更新与抽取待检测的体育比赛视频背景,将视频中的多个目标作为检测点,以此建立目标二维模型,对其坐标进行转换,根据坐标变换结果进行目标高分辨视觉重建,结合条件随机场模型设计体育比赛视频目标实时定位流程,实现目标实时定位。实验结果表明,该方法可以有效减少误检次数与漏检次数,提高定位效率,实际应用效果好。

Due to the randomness of target change,it is difficult to locate video target of sport game.Therefore,a real-time sport game video target location method based on conditional random field is proposed.We update and extract the sport game video background,take multiple targets in the video as detection points,so as to establish the target two-dimensional model,convert its coordinates,carry out target high-resolution visual reconstruction according to the coordinate transformation results,and design the real-time target positioning process of sports game video in combination with conditional random field model.The experimental results show that this method can effectively reduce the times of false detection and missed detection,improve the positioning efficiency,and has good practical application effect.