基于联合最佳基小波包的图像融合方法

基于联合最佳基小波包理论提出了一种新的图像融合方法。该方法首先将所有图像进行小波包分解,得到各自对应的四叉树,将各四叉树节点对应系数平方相加得到一个新的四叉树,利用此平方和四叉树搜索联合最佳基。将待融合图像基于联合最佳基分解,把分解系数加权处理得到融合系数,利用融合系数进行逆变换即得到融合图像。该方法对所有待融合图像的分解都是最佳分解,解决了先前最佳小波包融合方法只能最佳分解一方待融合图像的问题。将该方法与其他主要图像融合方法进行了比较分析,结果表明基于联合最佳小波包基的图像融合方法是非常有效的。

Image Fusion Method Based on Wavelet Packets with Joint Best Base

A newimage fusion method based on joint best basis wavelet packets is proposed. Images to be fused are first organized as a quadtrees to a specified depth. Then the squares of these quadtrees are summed into a sum- of- squares quadtree. The sum- of- squares quadtree is searched for best basis called joint best basis. Images to be fused are then decomposed with the joint best basis. The decomposition coefficients are dealt with weighted method and reconstructed into a fused image. Based on the joint best basis decomposition of all the images to fused is the best. However,for previous image fusion ways with best wavelet packets,the decomposition is best only for either images. The method is analyzed and compared with other major image fusion methods. The result validates that the newimage fusion method is very effective.