摘要
针对目标跟踪中的物体遮挡、光照影响、杂波扰动等问题,设计一种基于最小费用流建模的跟踪器.该跟踪器把整数规划与最小费用流模型相结合,将目标跟踪问题转变为可解的线性规划问题.与其他同类型跟踪器相比,该跟踪器具有更好的跟踪准确性.实验结果表明:运用最小费用流模型的跟踪器可以对复杂环境下的多个目标进行稳定跟踪,提升了跟踪的鲁棒性.
A tracker using the method of minimum cost flow was designed to solve the problems of object occlusion, lighting effects and clutter disturbance. This tracker combines integer programming with the minimum cost flow model, and converts the prob- lem of object tracking into a linear programming problem. Compared with other similar trackers, this tracker has a better tracking accu- racy. Simulation results show that the tracker in this paper has strong robustness and better tracking stability in complex environments.
出处
《工程科学学报》
EI
CAS
CSCD
北大核心
2015年第2期250-254,共5页
Chinese Journal of Engineering
基金
国家自然科学基金资助项目(61372090)
国家重点基础研究发展计划资助项目(2012CB821206)
北京市自然科学基金资助项目(4122037)
关键词
目标跟踪
建模
线性规划
机器视觉
target tracking
modeling
linear programming
computer vision
