摘要
为了在日趋复杂的空战环境中准确分析出目标的类型,以达到辅助决策之目的,采用自组织特征映射网络来对目标进行分类识别。首先提取影响目标识别的类特征,然后对其预处理。在此基础上建立SOM网络目标识别模型,并利用SOM网络算法实施无监督的自组织学习。在学习的过程中,通过不断调节网络节点间的权向量,来实现目标聚类。最后,通过仿真验证了该方法在目标分类识别中的可行性和实用性。
In order to accurately analysis types of targets in a complicated air combat environment and help decision-making, self organizing feature map is used in target classification and recognition. First of all, extract the sort-characters of target recognition, and then pre-treat these characters, on which basis, the SOM neural network model is created. At the same time, SOM-algorithm is used to non-supervised self-organizing study. In the process of self-organizing study, the weight vectors of neural network are continuously adjusted in order to achieving the cluster of targets. Numerical examples of simulation show that the method of target classification and recognition is valid and feasible.
出处
《火力与指挥控制》
CSCD
北大核心
2009年第1期101-104,共4页
Fire Control & Command Control
基金
“十一五”预研基金资助项目(KJ-050402011)
关键词
自组织特征映射网络
类特征
识别
权向量
聚类
self-organizing feature map, sort-character ,recognition, weight vector ,cluster