基于静态多小波变换的图像融合

Image Fusion Method Based on Stationary Multiwavelet Transform

综合了多小波分析和平移不变性质的优势,将多小波分析扩展到静态多小波的范畴,提出了一种基于静态多小波变换(SMWT)的图像融合方法。通过对源图像进行SMWT,根据变换系数的尺度内多子带联合窗口(CBWI)特性,实现多源图像的融合。该方法应用于一类多聚焦图像融合的仿真实验中,从视觉效果和信息量指标(互信息量和交叉熵)2个方面对融合图像进行主观评判和量化评价。结果表明,相比于传统小波域内的图像融合算法,该方法得到的融合结果具有更良好的视觉质量和更优的量化指标,体现出更强的融合性能。

A translation-invariant extension of multiwavelet transform called stationary multi-wavelet transform (SMWT) was proposed, which yields an overcomplete signal representation, and consequently incorporates the advantages of multiwavelet transform and translation-invariant property. Then, the image fusion techniques in the SMWT domain are studied, and a new image fusion algorithm based on cross-band-window-intrascale (CBWI) property in the SMWT domain is presented. Finally, the application of our proposed method to pixel level multi-focus images fusion problem is analyzed and compared with other image fusion methods. The simulation results indicate the method's superiority to the traditional wavelet-based image fusion method in terms of both visual quality and objective evaluation criteria.