摘要
针对传统交互式多模型算法实行正则滤波的单一化缺点,提出一种改进的跟踪算法。利用卡尔曼滤波匹配系统线性部分,粒子滤波匹配非线性部分,根据匹配深度判断目标遮挡程度,当目标被严重遮挡时,采用迭代的多级粒子滤波方法进行重采样,并结合卡尔曼滤波更新模型概率。实验结果表明,该算法实时性强,能提高模型滤波速度和目标状态的估计精度,缩短计算时间,解决跟踪过程中的遮挡问题。
Aiming at the singleness of implementing the regularized filter in interacting multiple model algorithm, an improved algorithm is proposed, in which Kalman Filter（KF） is used to match the linear part of the system and Particle Filter（PF） is used to match the non-linear part of the system, the degree of occlusion is determined according to the match extent. When the serious occlusion exists, the iterative multistage Particle Filter is exploited for re-sampling, then combined with Kalman Filter to update the model probability. Experimental results show that the proposed algorithm meets the real-time requirement, improves the speed of the model filter and the estimated accuracy of the object state, and reduces the computing time effectively. It solves the occlusion problem in the process of tracking.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第22期207-209,共3页
Computer Engineering
基金
甘肃省自然科学基金资助项目(0803RJZA025)
关键词
目标跟踪
卡尔曼滤波
粒子滤波
交互式多模型
遮挡
object tracking
Kalman Filter（KF）
Particle Filter（PF）
Interacting Multiple Model（IMM）
occlusion
