基于对比度的多分辨图像融合

Contrast-Based Multiresolution Image Fusion

本文给出一种基于对比度的多分辨图像融合算法 .首先 ,利用小波变换得到待融合图像的多分辨分析 ,同时得到图像的多分辨对比度序列 ;然后以对比度为判据 ,在图像的多分辨分析的相应各级上进行融合 ,得到融合图像的多分辨分析 ;最后 ,利用小波逆变换重构融合图像 .算法使用同一场景可见光图像和红外图像进行了验证 ,结果表明 ,融合图像完好地显示了源图像各自的信息 .

This paper introduces a contrast based multiresolution image fusion technique.Three steps can implement the method: First,the input multispectral images are decomposed by wavelet transform,meanwhile,the multiresolution contrast sequences of each input images can be obtained.Second,according to the contrast based criterion,the multiresolution analysis of the fused image can be obtained on the corresponding levels of the multiresolution analysis of the input images.Finally,the output image can be obtained through inverse wavelet transform.This algorithm is tested by fusing the visual and infrared images from same scence.The experiment shows that the fused image can preserve the details of the each input images successfully.