基于主成分分析的多传感器数据融合

针对多个传感器对某一特性指标进行测量实验的数据融合问题，根据多元统计理论，提出了一种基于主成分分析的融合方法。该方法把各传感器的测量数据作为一总体，定义总体的各主成分，利用测量值与主成分的相关关系，给出了各传感器的综合支持程度和数据融合公式。该方法不需要知道总体的分布和先验概率，避免定义距离矩阵和受主观因素作用的关系矩阵。应用实例验证了该方法的有效性和具有较强的抗干扰能力。

Fusion for multi-sensors data based on principal component analysis

Due to data fusion of multi-sensors experiment on some characteristic index,a new fusion method is proposed based on the principal component analysis by virtue of multivariate statistic theory.The method views the measured data of all sensors as a collectivity.After defining the each principal component for the collectivity,the synthesis support degrees of all sensors are given according to the relationship between the measured value and the principal component.The formula of data fusion is ob- tained.The method needn't to know the distribution and prior probability of the collectivity,and can avoid defining the distance matrix and the relationship matrix that is subjected to the subjective factors.The applied example proves that the method is effe- ctive and has the strong ability of anti-interference.