基于非采样提升小波变换的图像融合算法

针对传统提升小波变换的图像融合算法缺乏平移不变性以及区域能量融合准则存在的不足,提出了一种基于改进提升小波变换的图像融合新算法.该算法在提升小波变换原理的基础上,通过取消其奇偶分裂来获得具有平移不变性的非采样提升小波变换.对图像经此变换得到的低频部分设计出一种新的基于区域空间频率的加权和选择相结合的融合方法,高频部分采用一种基于边缘信息的加权融合策略.为验证所提出融合方法的有效性,对多聚焦图像进行了融合实验.实验表明,与传统的图像融合算法相比,该算法能更好地描述灰度的突变信息,获得含有丰富细节特征的融合图像.

Novel Image Fusion Algorithm Based on Unsubsampled Lifting Wavelet Transform

Traditional image fusion algorithms based on lifting wavelet transform have no trait of the shift invariance and the fusion rules based on regional energy have their own disadvantages,too.Thus,a novel new image fusion algorithm is proposed.Firstly,the algorithm cancels the odd and even split process of the lifting wavelet transform,and obtains the unsubsampled lifting wavelet transform with the trait of the shift invariance.Secondly,the original images are decomposed by the unsubsampled lifting wavelet.For ...