小波包变换的图像融合技术的研究  

Research of image fusion based on wavelet packet analysis

作　　者：穆静[1] 杜亚勤[1] 王长元[1] 张国华[1] 

机构地区：[1]西安工业学院计算机科学与工程学院,西安710032

出　　处：《西安工业学院学报》2005年第4期356-358,365,共4页Journal of Xi'an Institute of Technology

基　　金：陕西省教育厅科学计划专项科研计划项目(04JK241)

摘　　要：图像融合技术是把同一目标的多幅图像融合成一幅高质量的图像.给出了一种基于小波包变换的图像融合方法.对两幅图像分别进行小波包分解,得到两幅图像的小波包分解系数,进一步对小波包系数进行分析和处理.对两幅图像低频波段的逼近系数取均值,对高频段选取绝对值大的小波包系数,得到融合系数矩阵.对融合数据进行小波包反变换,得到两幅图像的融合图像.使用该方法对两幅图像进行了融合,得到的融合图像汲取了两幅图像的优势,视觉效果较好.The objective of image fusion is to form a high - quality image combining multiple images about the same object . The paper presents a method of image fusion based on wavelet packet transform. First, Two images are decomposed into wavelet packet coefficients respectively;the next step is to analyze and process the coefficients Approximate coefficients in low frequency band of two images and the average over absolute wavelet packet coefficients of high frequency brand form a fused coefficients matrix. The new image is obtained by taking the inverse wavelet packet transform of the fused wavelet packet coefficients. Experiment verifies the method：The result is suitable to the purpose of human visual perception.

关 键 词：分辨率分析 小波包分析 小波包系数 图像融合 

分 类 号：TN911.73[电子电信—通信与信息系统]