应用主成分分解(PCA)法的图像融合技术

PCA method used in image fusion

本文在非多尺度分解的框架下,针对像素级的图像融合,提出了一种基于PCA分解的图像融合算法。该算法利用主成分分解可以保留原数据中的主要信息这一特点,由源图像获取数据的协方差矩阵,协方差矩阵的特征值和特征向量,据此确定图像融合算法中的加权系数和最终融合图像。试验表明,应用该算法融合后的图像取得了满意的效果.

In the frame of non-multi-scale decomposition, a new fusion algorithm based on PCA is proposed for pixel-level image fusion. Covariance matrix, eigenvalue of covariance matrix and eigenvector of covariance matrix are got from source images because of the main information is preserved after PCA decomposition. Then we can get weight coefficient and fused image with the help of eigenvalue and eigenvector.The experimental results show that the better image quality is obtained in the proposed algorithm.