梅花形抽样不可分小波的多聚焦图像融合

A New Multi-Focus Image Fusion Based on Quincunx Sampling Non-Separable Wavelet

摘要:

Abstract: 本文提出的方法是针对离散小波变换(DWT)缺少移不变性和无下采样离散小波变换(UDWT)的高冗余及较大的计算量提出的一种方法，本文提出的方法是一种低冗余近似移不变的小波变换，通过介绍梅花形抽样与其它几种融合方法的比较，显现出梅花形抽样的多聚焦图像融合的优越性，即有较好的融合质量和相对较小的计算量。 This paper aims to provide a new approach to deal with the lack of shift invariance in Discrete Wavelet Transform (DWT), the high redundancy and the large amount of calculation involved in Undecimated-DWT (UDWT). In the paper, the author proposes a new method to achieve DWT with low redundancy and shows the advantages of multi-focus image fusion of quincunx sampling by comparing this sampling with other fusion methods. This novel method could arrive at the results with better fusion and reduce the amount of calculation in the process at the same time.