摘要
针对遥感图像,提出了一种遥感多光谱可见光图像与遥感高分辨率全色图像融合的小波分维算法。利用小波变换的方向选择性,对遥感多光谱的I分量和遥感全色图像进行小波分解,进而在不同子带中进行遥感图像融合,低频部分采用基于区域能量的融合算法;高频部分使用21×21的窗口加窗逐点计算分维作为权系数进行融合;最后将得到的灰度融合图像替代原来的I分量,进行反IHS变换得到最终的融合图像。实验结果表明,根据用分维进行数据融合方法确定两幅不同遥感原图像,在融合图像中所占的信息比例,可以有效地保留两幅原图像的边缘和纹理特征,避免融合图像平均化而出现的模糊现象,使得处理后的图像更容易识别,信息量增加。因此,文中提出的IHS小波分维融合算法是有效的,并取得良好的效果。
A more spectral visible light image and High-resolution Panchromatic image fusion algorithm based on the HIS, wavelet transformation and fractal dimension is developed for the domain of remote images. First, the remote panchromatic images and the component I of the remote multi-spectral image are decomposed to the domain of the wavelet transformation which has the advantage of good directions, and the remote image fusion is then implemen- ted in different subbands. The region energy fusion algorithm is adopted as fusion rules in low-pass subbands. In high-pass the fractal feature of every pixel is calculated by adding a size of 21 x21 of window, and then images of high-pass are merged with different fractal dimension of every pixel as weight. Finally, the stretched grayscale fused image replaces the original intensity component, and the final fused image is achieved. The experimental results show that, the use of the fractal feature to determine the scale of the two source image in the fused image can effectively hold the edge and textures of the two original images, and avoid the vague phenomenon in the distributions. The fused image is easier to identify and has more volume. Therefore, the algorithm of IHS and wavelet fractal raised in this paper is effective and can achieve better results.
出处
《电子科技》
2011年第1期4-8,共5页
Electronic Science and Technology
关键词
IHS变换
小波变换
分形维
图像融合
IHS transformation
wavelet transformation
fractal dimension
image fusion
