WorldView-2遥感图像融合新方法

摘要

New-style high resolution WorldView-2 satellite images pose challenges to the image fusion techniques. A new pansharpening method is proposed in this paper. First, 8-band multispectral imagery is resampled by nearest neighbor interpolation. According to the relative spectral responses between the multispectral band and the panchromatic band, a low spatial resolution panchromatic image is evaluated through multivariate linear regression. The spatial details are extracted from the original panchromatic image, and then injected into the component space of multispectral imagery. Finally, the pansharpened results are produced by employing inverse correspondence analysis transform. The experimental results show that the proposed method can obtain a better trade-off between the spatial resolution enhancement and the spectral information preservation compared to some existing methods.%新型高分辨率WorldView-2星载图像的出现给现有的图像融合技术带来了更大的挑战，该文提出了一种全色光和多光谱图像融合新方法。首先采用最近邻插值对多光谱图像重采样放大；然后结合WorldView-2各波段光谱响应特点利用多元线性回归构造出低分辨率全色光图像，通过对原始高分辨率全色光图像空间细节信息的提取并将其注入至多光谱图像的成分空间中；最后经对应分析反变换得到融合结果。实验结果表明，该方法在融合WorldView-2遥感图像时能够在提高空间分辨率和保持光谱信息两方面达到较好的平衡，优于现有的几种融合方法。