PCA-CCA在红外与可见光图像特征融合中的应用

针对红外与可见光图像特征级融合提出一种基于PCA-CCA的融合方法.分别提取红外与可见光图像的特征,由于当特征维数较高时,基于CCA方法的目标函数会面临协方差矩阵奇异的问题,无法进行求解,因此首先利用PCA方法进行降维,然后在低维空间中利用CCA方法求解融合特征.通过实验证明,本文的方法能够有效地提取融合特征,并且识别效果高于单一的CCA融合方法.

Feature-level Fusion of Infrared and Visible Images Based on Principal Component Analysis plus Canonical Correlation Analysis

A new fusion algorithm based on PCA （Principal Component Analysis） and CCA （Canonical Correlation Analysis） is proposed according to feature-level fusion of infrared and visible images.The features of infrared and visible images are abstracted respectively.When the feature dimension is higher,objective function based on the CCA method will face the problem of singular covariance matrix and can not be solved.PCA method is firstly used to reduce the dimension,and then CCA method is adopted to solve the fusion feature in the low-dimensional space.Simulation results show that the proposed algorithm could effectively extract the fusion feature,and the recognition effect is higher than single CCA identification method.