基于热传导模型的像素级遥感图像融合  

Pixel-level Remote Sensing Images Fusion Based on Heat Conduction Model

作　　者：许君一[1,2,3] 卢秀山[1,3] 卿熙宏[1,4] 姚继锋[1] 

机构地区：[1]山东科技大学测绘科学与工程学院,山东青岛266510 [2]山东省3S工程技术研究中心,山东青岛266510 [3]海岛(礁)测绘技术国家测绘局重点试验室,山东青岛266510 [4]山东科技大学资产处,山东青岛266510

出　　处：《测绘学报》2010年第6期566-571,578,共7页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金(30700107)

摘　　要：通过热传导方程给出一种像素级遥感图像融合模型和方法:①给出空间域内高分辨率图像与低分辨率图像之间的扩散关系,作为特例得到Brovey变换(brovey transform,BT);②给出图像融合与增强的统一表达式并得到基于亮度平衡的融合方法;③低分辨率多光谱图像的方差较小情形,指出基于方差的标准图像融合方法将会丢失高空间分辨率全色图像信息。试验表明,除了图像量化误差以外,所提议的方法不会丢失已知图像的空间分辨率和波谱信息。A novel pixel-level remote sensing image fusion based on heat conduction equation is proposed.The main contributions are as follows： ① the diffusion relationship between the high-resolution image and the low-resolution image are obtained in the space domain,where Brovey transform（BT） is one of its special cases;② a unified expression of pixel-level image fusion and image enhancement is obtained,and a brightness balanced-based image fusion is obtained by using the expression;③ the disadvantage that the standard variance-based image fusion method will result in the loss of high spatial resolution panchromatic image information,for smaller variance of low-resolution multispectral images is pointed out.Experimental results show that this approach does not lose the image spatial resolution and spectral information and outperforms the existing method.

关 键 词：图像融合 遥感 HSI 多光谱图像 热传导方程 

分 类 号：P237[天文地球—摄影测量与遥感]