摘要
针对经典卡尔曼滤波器在滤波数学模型与实际过程的数学规律不匹配、滤波特性较差的情况,提出利用强跟踪卡尔曼滤波器对视频序列图像中的运动目标进行跟踪。该方法是在经典卡尔曼滤波递推公式中的一步验前误差方差阵中引入可在线计算的时变渐消矩阵,从而调节增益K,使之能够不断变化,保证对新息序列的自适应调节,使状态滤波更准确。实验结果表明,较之经典卡尔曼滤波,该方法具有对运动目标更强的跟踪能力,跟踪精度更高,均方误差更小。
For the poor situation of the classical Kalman filter in filtering of mathematical models and the actual process of mathematical laws mismatch as well as the poorly filtering characteristics,one method of strong tracking Kalman filter in the video image sequences to track the moving target is proposed.This method is based on the introduction of the time-varying fading matrix which can be calculated on-line in the priori error covariance matrix of classical Kalman filter’s recurrence formula,which can enable the gain to time-changing and ensure the adaptive adjustment of the new sequence,as well as make the state filtering more accurate.Experimental results show that the strong tracking Kalman filter in the video moving target has a greater ability,higher accuracy,and the smaller mean square error compared to classical Kalman filter.
出处
《计算机工程与应用》
CSCD
北大核心
2011年第11期128-131,166,共5页
Computer Engineering and Applications
基金
浙江省教育厅科研计划项目No.Y200804700~~
关键词
强跟踪卡尔曼滤波
视频图像
目标跟踪
strong tracking Kalman filter
video image
target tracking
