基于离散多小波变换的遥感图像融合方法

以不同分辨率的遥感图像为对象，基于Chui—Lian（CL）离散多小波变换的特性，提出了一种新的图像融合方法．该方法将两幅不同的源图像分别进行预处理和多小波分解得到各个分解图像，然后对分解图像分别采用基于区域特征的融合方法，得到混合的分解系数，通过多小波重构和后处理算法从而获得融合图像．该方法能够为图像融合提供一种比传统的小波变换更加精确的融合方法．实验结果证明采用这种方法可以得到更好的融合效果，不仅能够完好地显示源图像各自的信息，而且能更好地将源图像的细节融合在一起．

Remote sensing image fusion by using discrete multiwavelet transform

The object of research in this paper is remote sensing images with different resolution.A new algorithm of image fusion was presented, after the characteristics of Chui-Lian(CL) discrete multiwavelet transform being researched.In this algorithm,two input images are firstly pre-processed and decomposed by CL multiwavelet,respectively.Then,two series of subband images can be gotten.A new fusion rule was used to merge coefficient by using region feature of subband images.After multiwavelet reconstruction and post-processing,a new fused image was gained.This algorithm can offer a more precise way for image fusion than traditional wavelet transform.Comparing with other image fusion algorithm and appraising their performance,it was concluded that this algorithm can fuse details of input images successfully,and it can perfectly display information of each input image.The results of the experiment show that this image fusion algorithm can get more satisfactory result than that of other algorithms.