Kalman Mean Shift 3D残差跟踪算法

Kalman Mean Shift 3D Difference Tracking Algorithm

本文针对Mean Shift对快速运动物体跟踪效果差,每帧中目标的起始位置选取不够理想和存在的大比例遮挡问题,提出了一种新的算法,即基于三维立体空间的Kalman Mean Shift 3D残差判断跟踪算法。实验结果表明该算法能够实现快速运动目标的跟踪以及能够解决场景中运动物体大比例遮挡的问题。

Due to the bad effect of tracking fast-moving objects and the problem of the unideal initial position and the severe occlusion based on Mean shift algorithm, this paper presents a new algorithm named Kalman Mean Shift 3D difference algorithm. The simulation shows that the proposed algorithm can track the fast moving objects effectively and can solve the problem when the target is almost sheltered.