基于Kalman滤波的CAMShift目标跟踪算法

CAMShift target tracking algorithm based on Kalman filtering

针对传统的CAMShift算法跟踪目标丢失的问题,提出一种基于Kalman滤波的CAMShift目标跟踪算法。为验证改进后算法,可用安装摄像头的农田作业设备采集图像,并对图像中的特定目标进行跟踪。该算法用Kalman滤波器预测下一帧特定目标的位置,统计候选目标的直方图并进行反向投影,将得到的色彩概率分布图跟踪特定目标的特征。实验表明,改进后的算法在目标快速运动的情况下仍然取得较好的跟踪效果,具有较好的稳定性。

Aiming at the problem of tracking target loss by the traditional CAMShift algorithm,a Kalman filtering based CAMShift target tracking algorithm is proposed.To validate the improved algorithm,images can be acquired using a farmland-mounted farm equipment and tracking specific targets in the image.The algorithm uses the Kalman filter to predict the position of a specific target in the next frame,counts the histogram of the candidate target and performs back projection,and traces the obtained color probability distribution map to the characteristics of the specific target.Experiments show that the improved algorithm still achieves better tracking effect in the case of fast target motion and has better stability.