基于粗糙集的炮兵地面目标识别特征量选取方法  

Method of Artillery Ground Target—recognition Features Selection Based on RS

作　　者：张敬忠[1] 陈栋[1] 窦维江[2] 

机构地区：[1]解放军炮兵学院,安徽合肥230031 [2]江南大学信息学院,江苏无锡510023

出　　处：《计算机测量与控制》2008年第6期815-818,共4页Computer Measurement &Control

摘　　要：地面目标识别是炮兵侦察的主要内容之一,其关键在于目标特征量的选取;文章在Kohonen的自组织特征映射(SOM)方法基础上,同时结合概率分类方法,提出了一种最优分类器方法,并将其用于炮兵地面目标特征量的量化;采用粗糙集理论,实现了对经过量化后的目标特征量进行简化和优化,指导目标特征量的选取,为后续的使用学习机器更快更准确的识别目标提供重要的前提条件。Target-recognition is the Primary part of artillery reconnaissance and target features selection is the importantance. Based on Kohonen＇s SOM method, an optimal classifier is proposed to quantify target features with probability classification. By rough set theory the article instructs the target feature selection to simplify and optimize the target features. After that learning machine can recognize the target faster and exactlier based on the previous target features.

关 键 词：目标识别 粗糙集 目标特征量 特征量选取 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]