Mean-Shift结合Kalman滤波的车辆目标跟踪算法

Algorithm of Mean-Shift Combined with Kalman Filter Vehicle Target Tracking

近年来,在智能交通系统与车辆辅助驾驶系统中,作为一个非常重要并具有挑战性的课题——车辆检测与跟踪技术,对其进行了大量的研究。尤其在智能交通系统中,必须了解路面其他车辆的位置才能确保行驶安全,而这些系统的基础——稳定可靠的车辆检测与跟踪,就显得尤为重要了。因此,深入研究运动车辆目标检测与跟踪算法,考虑到系统的实时性,系统采用帧间差分的方法对车辆目标进行检测,建立的Mean-Shift结合Kalman滤波的车辆目标跟踪模型,对车辆目标进行跟踪,提高了跟踪准确率与实时性。

In the past years,in intelligent transportation systems and driver assistance system,as an important and challenge issue,vehicle detection and tracking has been broadly investigated. because of the importance of helping driver assistance system,determining the position of other vehicles on the high speed road is the basic step. Thus,In these systems,it is necessary of robust and reliable vehicle detection and tracking. This paper studied vehicle detecting and tracking algorithms in depth,considering system real time,using the difference between neighborhood images to detect vehicle,a model based on Mean-Shift and Kalman filter to track vehicle,improving accuracy.