摘要
以智能交通系统中运动目标跟踪为背景,研究采用均值漂移算法,实现对快速移动的单目标事物准确实时跟踪。首先,分析了均值漂移算法的基本原理。其次,分析了均值漂移算法在目标跟踪中存在的问题及解决方法:针对目标遮挡情况下目标跟踪的准确性问题,采用将均值漂移算法与卡尔曼滤波算法相结合的自适应均值漂移目标跟踪算法;针对大小不变的带宽窗口可能导致跟踪丢失的问题,采用了自动调节带宽窗口的方法。最后进行了仿真实验。
Fast moving object was accurately tracked based on the intelligent transportation system. Firstly, the principle of mean shift algorithm was analyzed. Then, the mean shift algorithm in target tracking problems and solutions were analyzed： target under occlusion tracking accuracy, the mean shift algorithm and Kalman algorithm combining adaptive mean shift tracking algorithm; size constant bandwidth window leading to loss of tracking problem; automatic adjusting the bandwidth of the window method. Finally, through the simulation experiments, the tracking effects were compared.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2012年第9期1896-1899,共4页
Journal of System Simulation
基金
河北省自然基金(F2012208004)
关键词
均值漂移
目标跟踪
智能交通
卡尔曼滤波
mean shift
object tracking
intelligent transportation
Kalman
