基于Sobel算子和局部能量的图像融合新算法  

Novel Image Fusion Algorithm Based on Sobel Operator and Region Energy

作　　者：魏世超[1] 段先华[1] 夏加星[1] 

机构地区：[1]江苏科技大学计算机科学与工程学院,江苏镇江212003

出　　处：《计算机技术与发展》2012年第4期61-64,共4页Computer Technology and Development

基　　金：江苏省自然科学基金(BK2008411)

摘　　要：文中提出一种新的基于小波变换的图像融合算法,以提高图像融合质量。首先应用小波变换将图像分为高频和低频部分,再对小波的高频和低频系数采用融合算法进行处理,最终将处理后的高频和低频小波系数进行融合。文中对低频系数采用基于Sobel算子的方法,有效保留边缘特征。对高频系数采用基于局部能量取大准则,有效保留高频的细节和区域特征。为了抑制噪声和控制图像的不稳定性,最后再对低频和高频系数的选择结果进行邻域窗口的一致性检验与调整。与传统金字塔方法和经典小波变换相比较,实验结果表明此方法融合效果较好。An new image fusion algorithm based on wavelet transform was proposed to promote the image fusion quality.Decompose the image into high-frequency part and low-frequency part by wavelet transform firstly,then process the high-frequency part and low-frequency part using fusion algorithm,finally fuse the processed high-frequency coefficients and low-frequency coefficients.When choosing the low-frequency coefficient,the method which used is based on Sobel operator,retain the edge characteristics effectively.The high-frequency coefficients are selected with region energy which is maximum,so that retain details and area characteristics effectively.In order to overcome the presence of noise and control of image instability,all the coefficients are subsequently performed by a window-based consistency verification process.Compared with traditional pyramid method and classical wavelet transform,the experimental results show that the fusion rule of proposed method is effective.

关 键 词：图像融合 小波变换 SOBEL算子 局部能量 一致性检验 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]