一种基于自适应权值灰关联分析的毫米波/红外融合识别方法  

A Method of MMW/IR Fusion for Target Recognition Based on Gray Relational Analysis with Adaptive Weights

作　　者：郭军[1] 赵和鹏[2] 朱长仁[1] 

机构地区：[1]国防科学技术大学ATR重点实验室 [2]海军91635部队

出　　处：《计算机工程与科学》2010年第2期146-149,共4页Computer Engineering & Science

基　　金：国防科技重点实验室基金资助项目(9140c8004010611)

摘　　要：本文提出了一种基于自适应权值灰关联分析和DS理论的毫米波/红外融合目标识别方法。首先对单个传感器提取的特征计算灰关联度作为基本概率赋值,再用DS理论进行融合得到识别结果。在灰关联度计算中,本文提出了使用熵值分析法自适应确定不同特征的权值,提高了系统性能。基于实测数据的实验结果验证了本文的有效性。A method of MMW/IR fusion for target recognition based on gray relational analysis with adaptive weights and the DS theory is presented. Firstly, the gray relational grade of each sensor is computed and treated as the basic probability assignment function. Then, the fusion recognition result is computed by the DS theory. At the step of gray relational grade calculation, an algorithm of entropy analysis is introduced to adaptively determine each weight, and then the system performance is improved. The experimental results based on real data show that the proposed method is effective.

关 键 词：信息融合 目标识别 灰关联分析 自适应权值 熵值分析 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]