人体运动目标跟踪的滞后性问题研究

Research Hysteresis Problems of Human Motion Tracking

研究人体运动目标跟踪问题,由于图像目标跟踪实时性差,在摄像机与被跟踪物体同时运动的情况下,被跟踪人体走路速度过快,两者速度不匹配时,运动的人体存在着被跟踪的滞后性问题。为解决上述问题,提出了一种改进的粒子跟踪方法。在人体跟踪的过程中,可以根据实际跟踪情况在线减少粒子的数目,进而减少运算时间,使得算法的运算速度可以根据被跟踪人体的运动速度实时调节,形成速度最佳匹配。`仿真表明,改进的算法很好地解决了被跟踪运动目标的滞后性的问题,跟踪效果明显改善。

To study the human movement target tracking,aiming at the simultaneously movement of camera and tracked objects.Human body moves too fast,and the camera speed does not match.This paper proposes an improved particle tracing method.In human body tracking process,,acording to the actual tracking situation,the method can dynamicly reduce the number of particles,and then reduce the time cost,make the algorithm speed be adjusted on time,and form optimal speed matching.Experiments show that the improved algorithm can solve the problem of tracking moving targets synchronously,and the tracking performance is improved obviously.