多传感器数据的Fisher判别法  

Method of Fisher discrimination for multi-sensors data fusion

作　　者：万树平[1] 

机构地区：[1]江西财经大学信息管理学院,江西南昌330013

出　　处：《传感器与微系统》2006年第8期61-62,65,共3页Transducer and Microsystem Technologies

摘　　要：在多传感器信息融合过程中,有时要利用多传感器对某一对象进行不同方面观测,而数据融合的目的是要对研究对象进行类型识别,根据判别分析的含义,给出了一种F isher判别法。该方法不需要知道总体的分布类型,可以避免丢失重要的极端数据,使结论更符合实际。During information fusion of multi-sensors,sometime object should is observed from different aspects, the goal of data fusion is discriminating the belonging type of the research object. Based on the connotation for discriminant analysis, a new method of Fisher discrimination for multi-sensors data fusion is proposed. The method needn＇t know the distribution function and can avoid losing the important extreme data and make the conclusion more suitable for the practice.

关 键 词：多传感器 数据融合 特征指标 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]