多源遥感图像融合方法研究

摘要

研究多源遥感图像的融合技术,针对不同传感器获取的遥感图像像素信息有很大差异.当所要融合的图像是多源遥感图像时,应提取多图像的有效信息,组合出高质量的图像.传统IHS图像融合方法无法避免多源图像像素不匹配带来的有效像素丢失,造成融合图像模糊,清晰度不高的问题.提出一种基于Contourlet变换的遥感图像融合方法,通过对图像进行Contourlet变换后提取各源图像的特征信息,并计算提取特征所包含的信息量,选取高信息景的特征进行融合,最后通过进行Contourlet逆变换即得到多源融合图像,利用信息量融合配准的方法就避免了直接对不匹配像素运算而造成的有效像素丢失.实验证明,改进的优化融合方法能够保留图像的有效信息,得到高清晰度的融合图像.%Research multi - source remote sensing image fusion methods. The remote sensing pixels spectrum information from different sensors has very big differences. When the fusion images were multi - source remote sensing images, traditional HIS image fusion method cannot avoid the effective pixels lost because multiple source pixels do not match, causing the problem of fuzzy fusion images. This paper presented a contourlet algorithm based on the remote sensing image fusion method. The characteristics information of source images were extracted after the image contourlet transformation. Then the information content contained in the extracted features was calculated, and the characteristics with high information content were selected for fusion. Finally, through the inverse contourlet transformation , the multi - source fusion image was produced. This information fusion registration method can avoid the pixel lost when the unmatched pixels are operated. Experiments show that the optimal fusion method effective information can keep images, get high quality fusion image.