Abstract: This paper explains the theory and method of image fusion based on principal component analysis of the Laplacian pyramid. The Laplacian image fusion scheme begins by constructing Laplacian pyramids for each source image, and then for the high frequency part it uses the Principal Component Analysis（PCA） fusion method, for the low frequency part it uses the average gradient method. Finally, the end fused image is obtained by inverse Laplacian pyramid transform. By analyzing the fusion image with visible and infrared image, and image fusion of different focal images, the experimental results show that this algorithm can produce high-contrast fusion images that are clearly more appealing and have greater useful information content than the PCA and the Laplace image fusion．

Key words: image fusion, Laplacian pyramid, Principal Component Analysis（PCA）, average gradient

摘要： 阐述了基于主元分析的拉普拉斯金字塔图像融合的原理和方法：对原图像分别进行拉普拉斯金字塔分解，分别对高频部分采用主元分析（PCA）法融合，对低频部分采用平均梯度法进行融合，对拉普拉斯金字塔做反变换得到最终的融合图像。通过对可见光与红外图像的融合，以及对不同焦距图像融合的结果分析，该算法比单纯的PCA和拉普拉斯图像融合能得到具有更多有用信息的高对比度的融合图像。

关键词: 图像融合, 拉普拉斯金字塔, 主元分析, 平均梯度