基于小波分解的融合图像评价方法  

A Wavelet-Decomposition-Based Fusion Performance Measure

作　　者：裴红周[1] 陈炜[1] 毛士艺[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100083

出　　处：《信号处理》2006年第2期268-271,共4页Journal of Signal Processing

摘　　要：本文提出了一种新的基于小波多尺度分解的图像质量评价方法,其基本思想是首先对源图像和融合图像进行小波多尺度分解;其次,在小波域计算源图像与融合图像的相关性;最后根据小波系数的重要性进行加权得出评价指数。仿真表明:该方法的算法简单,计算量小,评价结果符合人眼的视觉特性,可用于多传感器图像融合效果评价。A novel image fusion performance measure based on wavelet multiscale decomposition is presented in this paper. The basic idea is to perform a wavelet multiscale decomposition of each source image and each resulting fused image first, then describe the relativity between the source images and the fused image in the wavelet region, finally the result to evaluate the fusion image quality is obtained by normalizing the relativity according to the visual importance of the wavelet coefficients. The experimental results show that the new algorithm is simple and efficient and performs well in accordance with human subjective vision. It can be applied to assess the multisensor image fusion quality.

关 键 词：图像融合 小波分解 图像质量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TS807[自动化与计算机技术—计算机科学与技术]