两种基于2DPCA图像融合算法的性能研究  

Performance of two kinds of 2DPCA algorithms in image fusion

作　　者：吴学明[1] 杨武年[1] 

机构地区：[1]成都理工大学国土资源信息技术与应用国土资源部重点实验室,成都610059

出　　处：《测绘科学》2009年第3期127-129,共3页Science of Surveying and Mapping

摘　　要：基于2DPCA的图像融合算法是一种新近提出的图像融合技术,它是通过图像矩阵右乘投影矩阵来实现2DPCA分析的。因而,这种算法能够有效利用图像的二维结构信息,具有良好的性能,不仅能够提高融合图像的空间分辨率,而且使光谱信息得到较好保持。提出了一种通过图像矩阵左乘最优投影矩阵来实现2DPCA分析的图像融合算法,即L2DPCA图像融合算法,并将2DPCA算法和L2DPCA算法与经典的BROVEY、PCA和H IS三种图像融合算法进行了对比研究。结果表明:2DPCA算法和L2DPCA算法的性能明显优于这三种经典的算法。2DPCA-based image fusion algorithm is a newly proposed technique in image fusion. 2DPCA（ Two Dimensional Principal Component Analysis） is directly performed on the image matrices by multiplying the optimal projection matrix at the right side of the muhispectral images. The advantage of this technique is that it can not only effectively utilize the structural information of the multispectral images, but also improve the spatial resolution and preserve the spectral information. The paper compared the performance of 2DPCA algorithm and the proposed L2DPCA algorithm with that of HIS, PCA and Brovey techniques. Experimental results showed that 2DPCA and L2DPCA outperform the other three classical algorithms.

关 键 词：图像融合 遥感 2DPCA L2DPCA 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]