海战场多目标聚类-匹配分群研究  

The Research on Multi-Target Grouping on the Sea Battlefield Based on the Clustering-Matching Method

作　　者：刘吉军[1] 李刚[1] 吴言凤[1] 

机构地区：[1]海军潜艇学院,山东青岛266042

出　　处：《四川兵工学报》2013年第2期104-106,120,共4页Journal of Sichuan Ordnance

摘　　要：群体已取代单一实体成为兵力执行海上作战任务的主要样式,多目标分群因而成为兵力层态势信息的形成前提亟待解决,从海战场作战群体的特点入手,使用聚类与匹配相结合的方式构建分群模型;首先从实体层目标信息中选取特征矢量与聚类测度,运用改进ISODATA聚类与辅助信息识别形成战场空间群,然后结合军事先验知识通过模板匹配进一步将空间群聚合为功能群等高层级群体;该模型能够满足海战场不同层级群体识别的需要,并可为其他背景下的聚类分群提供借鉴。Group has become the main pattern of force performing marine operational mission instead of single entity, so the multi-target grouping, that becomes the precondition of force level situation informa- tion forming, is the problem waiting for solution. Starting with the characteristic of operational groups on the sea battlefield, grouping model is constructed by means of clustering and matching. Firstly, character vector and cluster measure are selected from entity level target information, and battlefield space-group is formed using improved ISODATA and assistant information identification. Secondly, space-group is aggre- gated to function-group and other high level group with template matching which combines a prior military knowledge. The model can meet the necessary of marine different level group identification, and use for reference for grouping under other background.

关 键 词：多目标分群 动态聚类 改进ISODATA 模板匹配 

分 类 号：E911[军事]