多传感器目标识别的改进灰关联度法  

Method of improved grey relational degree for multi-sensor target recognition

作　　者：万树平[1] 

机构地区：[1]江西财经大学信息管理学院,南昌330013

出　　处：《计算机工程与应用》2009年第24期25-27,共3页Computer Engineering and Applications

基　　金：国家自然科学基金No.10626029;江西省自然科学基金No.0611082~~

摘　　要：针对具有多个特征指标的多传感器目标识别问题,提出了一种新的融合方法。该方法利用置信距离改进了灰关联系数,通过求解各属性的置信距离的偏差平方之和最大的优化问题,获得属性的权重,从而给出目标识别方法。克服了特征权重选取的主观性,提高了目标识别结果的准确性和可信度。仿真实例验证了方法的有效性。Aiming at the target recognition problem of multi-sensor with multiple characteristic indexes,a new fusion method data is proposed.The method uses the distance of confidence to improve the grey relational coefficient.By solving the optimal progranming of maximizing the total square deviation of the confidence distance for all attributes,the weights of attributes are derived. Hence,the method of object recognition is given.The method can overcome the subjectivity of the weights acquisition of attributes,and improve the accuracy and trustworthy degree of target recognition.The simulated example proves the effectiveness of the method.

关 键 词：多传感器 数据融合 目标识别 灰关联度 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]