结合Camshift和Kalman预测的运动目标跟踪

Camshift and Kalman Predicting Based on Moving Target Tracking

针对单一的CamShift跟踪算法在目标发生遮挡时非常容易致使跟踪目标失败的问题,本文提出了一种基于CamShift和Kalman预测的跟踪算法。首先,采用帧间差分阈值法来快速、精确地检测和提取出运动目标;然后,通过在CamShift算法中使用运动目标的颜色特征,在图像序列中找到运动目标的所在位置和大小;最后,使用Kalman滤波预测目标的位置,进而有效地解决了背景中大面积相同颜色的干扰和目标部分被遮挡等问题。用无线遥控车完成了运动目标的跟踪实验,实验证明结合CamShift算法和Kalman预测滤波能实时、准确地跟踪目标。

This paper presents a tracking algorithm based on the CamShift and Kalman prediction which was proposed to solve the poor tracking ability problem in occlusions just using single Camshift. Firstly, an inter-frame difference threshold method is used to achieve the detection and extraction of the target rapidly and accurately. Secondly, the color characteristics of moving objects in CamShift algorithm are used to find its location and size in image sequences. Finally, to deal with the object occlusion by other objects which have similar color, a Kalman filter is used to track the object region centroid. We use a wireless remote-controlled car to achieve a moving target tracking experiment. The experimental results denote that combining the Cam- Shift algorithm and the Kalman prediction filtering completes a real-time and accurate target tracking.