中文摘要：

结合区域分割理论,提出了一种基于Shearlet变换的遥感图像融合算法。首先通过色度H-亮度I-饱和度S（HIS）变换对多光谱图像进行分解,将I分量与全色图像进行Shearlet变换,得到低频、高频信息图;然后对低频信息图进行基于灰度的区域分割,两图像的低频部分使用改进的加权融合算法改善融合图像轮廓模糊问题,以区域匹配度作为融合规则,高频分量采用区域清晰比作为融合规则,得到更多的细节信息;最后通过HIS逆变换得到融合图像。实验结果表明,本文算法所得融合图像在有效地保持了多光谱图像光谱信息的同时,提高了空间细节信息。

英文摘要：

A remote sensing image fusion method based on Shearlet transform and region segmentation is proposed. First,the multi-spectral image is transformed into hue-intensity-saturation （HIS） color space. The intensity component of the mnlti-spectral image and the panchromatic image are decomposed by Shearlet transform to obtain low frequency and high frequency information images. Second,the low frequency parts of multi-spectral image are segmented into regions by means of thresholds method. The results show that the proposed algorithm can not only retain the spectrum information source by the multispectral image, but also improve the spatial detail information.