基于边缘检测与归一化相关矩的图像融合方法  

Image fusion based on wavelet edge detection and normalized correlation moment

作　　者：高国荣[1] 李文敏[1] 潘琼[1] 

机构地区：[1]西北农林科技大学理学院数学系,陕西杨凌712100

出　　处：《计算机工程与应用》2008年第27期194-196,共3页Computer Engineering and Applications

基　　金：西北农林科技大学人才基金

摘　　要：常见的基于小波变换的图像融合方法只是从小波系数的某一特性来考虑小波系数的选取,因而融合结果往往只提升了融合图像的某一特性。从数据融合的角度,提出了一种有效的融合算法,该方法将基于小波变换的图像边缘检测与归一化相关矩相结合并应用于图像的融合中。实验表明,与单一的融合规则相比,该方法能有效提高融合图像的质量。As common image fusion schemes based on wavelet often consider a certain aspects of task when fusing wavelet coefficient of the source images,it is most possible that these methods often enhance a single property of the fused image.From the point of data fusion,a new fusion scheme is proposed.This new method band edge detection based on wavelet and the normalized correlation moment together.Experimental results show that this method is effective compared with the single fusion method in the quality of fusion result.

关 键 词：图像融合 小波变换 小波金字塔 归一化相关矩 

分 类 号：TN911.73[电子电信—通信与信息系统]