基于光谱范围的IHS改进融合方法研究

Research of Improved IHS Image Fusion Method Based on Spectral Range

针对Intensity-Hue-Saturation(IHS)融合算法对高分二号卫星多光谱(MS)影像和全色(PAN)影像融合时所产生的光谱失真现象,提出了一种基于光谱范围的IHS融合改进方法(SIHS).该方法首先将多光谱影像进行重采样,使其与全色影像大小相同,然后使用光谱范围对重采样后的多光谱影像各波段加权平均获得亮度分量I,再对全色影像和I进行直方图匹配,最后使用快速IHS融合算法得到最终融合影像.通过实验,将该方法的融合结果与FIHS、Gram-Schmidt和NNDiffuse融合结果进行主客观评价.结果表明该方法在对光谱信息保留上优于其他融合算法,可以更好地支持影像信息的分类和检测.

In order to achieve high-quality remote sensing imagery, the Intensity-Hue-Saturation （IHS）fusion algorithm was often utilized to generate the synthetic imagery by fusing multispectral （MS） imagery and panchromatic （Pan） imagery. However, the spectral distortion phenomenon was easily generated by using IHS algorithm to fuse MS imagery and Pan imagery, which were obtained by GF-2 satellite. To solve this problem, the paper proposes an improved IHS fusion method based on spectral range. This method will first resample the multispectral image to fit the size of the panchromatic image, and then use the spectral range of multispectral image as weight to gain intensity component I, further match the panchromatic image with I through histogram, finally get the fusion image by FIHS fusion algorithm. By comparing our proposed algorithm with FIHS, Gram-Schmidt and NNDiffuse in objective and subjective evaluation, the experimental results demonstrate that the proposed algorithm has satisfactory performance in preserving spectral information. This algorithm is helpful for the research such as classification and detection of image information.