摘要
提出了基于运动目标历史速度和历史运动曲线的改进的粒子滤波器设计方案。为了解决运动中的遮挡问题,在粒子评价过程中引入"运动动量因子"来保持在高速运动下原来方向粒子的健壮性,提高了粒子跟踪的准确度。相应的为运动目标建立颜色模型,进一步增加了跟踪的准确性。实验结果表明,跟踪的效果优于基本粒子滤波器。
By analyzing the velocity of motion object, the basic particle filter based the velocity and curve of history images is improved. To solve the occlusion problem, adding the motion inertia factor to the process keeps haleness of the particles and improves the tracking veracity. And, modeling color of the motion object improves the veracity of tracking motion object. The experiment shows the result is better than basic particle filter.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第15期3968-3971,共4页
Computer Engineering and Design
关键词
目标跟踪
卡尔曼滤波
粒子滤波
粒子退化
速度
遮挡
object tracking
Kalmanfilter
particle filter
particle degenerating
velocity
occlusion