基于提升小波变换的图像融合算法研究  

Research of image fusion algorithm based on LWT

作　　者：范璐璐[1] 安成斌[1] 

机构地区：[1]华北光电技术研究所,北京100015

出　　处：《激光与红外》2012年第9期1076-1079,共4页Laser & Infrared

摘　　要：介绍了将来自不同传感器的图像进行融合的改进方法。涵盖了配准,图像分解,融合,图像重建四个主要步骤,并进行了仿真,最后用评价函数对不同方法得到的图像进行评价。在融合准则上提出了一些新的处理方式。试验结果表明,融合后的图像确实能够汲取两幅原图像的各自不同特点,达到了较好的效果。本文还针对工程应用对算法进行了优化。Image fusion from different sensors is described in this article. We use mid-wave and long-wave IR images with different size, different angle but focusing on the same scene as the source images. After rotating, cropping, and scaling, they are matched. And then we use lifting wavelet transform to decompose and reconstruct the images. On fusion rules, we tried a new way on the low frequency parts. Additionally, quantitative evaluation of fusion performance is presented. The experiment result shows that the fusion image contains the different feature of the two source images. The algorithm has been optimized and applied in practical project.

关 键 词：图像处理 图像融合 提升小波 配准 

分 类 号：TN911.73[电子电信—通信与信息系统]