基于多尺度图像融合算法研究

Research on Image Fusion Algorithm Based on Multiple Scales

摘要: 针对传统基于小波变换的图像融合方法不具有平移不变性的缺点，本文进行了创新，并提出了一种结合小波变换和自适应分块的多聚焦图像融合算法。通过与传统小波变换的对比来凸显多聚焦图像融合算法的优势，该算法既可以一定程度克服小波变换不具有平移不变性从而对非严格配准区域融合效果差的缺陷，又可以有效抑制直接在空域进行分块融合易产生块效应的缺点。

Abstract: Aiming at the disadvantage that the traditional Image fusion method based on wavelet transform does not have translation invariance, this paper innovates and proposes a multi-focus image fusion algorithm combining wavelet transform and adaptive block. By comparing with the traditional wavelet transform, the advantages of multi-focus image fusion algorithm are highlighted. The algorithm can not only overcome the defect that wavelet transform does not have translation invariance, so the fusion effect of nonstrict registration region is poor, but also can effectively suppress the block effect of block fusion directly in spatial domain.