摘要
针对群目标跟踪中群起始的特殊性,阐述了群起始的作用与意义,描述了群起始现有的K-方法和集群引晶方法的基本方法、步骤及存在的问题。在此基础上,给出了群起始的定义,提出了基于群目标几何中心的起始算法。该算法首先根据监视区域内多目标回波的空间距离进行分群检测预处理,在完成分群检测后再求解每个群目标的几何中心,然后确定该门限区域内的目标回波相对于几何中心的权值,最后构建出等效量测即群目标质心,从而实现了群目标快速、稳定的群起始。仿真结果证明了该算法的有效性。
Group-target tracking is a special problem compared with ordinary mupti-target tracking. So, the purpose and meaning of group-initialization are set forth for group-target tracking. The steps and basic meaning of K-meanings and cluster-seeding methods for group-initialization are described. The group-initialization is defined, and the group-initialization method based on the center of geometry is proposed. Firstly, all echoes in the surveillance space are made sure by the divided group detection. Secondly, the center of geometry of every subgroup is calculated. Thirdly, the weight of each echo in the association area is computed. At last, the equivalent echo of the group-target and its centroid are found, and the group-initialization is made. Simulation shows the efficiency of the method of group-initialization for group targets.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第2期269-272,共4页
Systems Engineering and Electronics
关键词
群目标
群起始
K-方法
集群引晶
群目标几何中心
群目标质心
group-target
group-initialization
K-means
cluster-seeding method
group-target center of geometry
group-target centroid