àtrous小波-NSCT遥感图像融合  

Remote sensing image fusion based on àtrous-NSCT transform

作　　者：杨建慧[1] 白超[1] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004

出　　处：《计算机工程与应用》2013年第5期170-173,共4页Computer Engineering and Applications

摘　　要：在IHS空间变换基础上,利用非下采样Contourlet变换中具有多方向性、平移不变性的非下采样方向滤波器组,并结合àtrous小波变换,实现了一种基于àtrous小波-NSCT变换的遥感多光谱图像和全色图像融合方法。针对传统的细节注入法融合规则会引起较大光谱失真,对变换所得的高频分量采用特征量积表达局部图像细节特征,依据决策因子阈值抽取全色图像细节信息,将经反方向滤波得到的有效高频细节面附加给多光谱图像分量。实验结果表明,在多种不同性能指标评价下,该算法得到的光谱扭曲最小,相关系数最大,同时空间细节质量也得到很好改善。A fusion scheme for multispectral image and panchromatic image is proposed, based on Intensity, Hue, and Saturation （IHS） transform and Nonsubsampled Directional Filter Bank （NSDFB） included in nonsubsampled contourlet transform which provides a full shift-invariant directional multi-resolution trait, combined with the superiority of fitrous transform. In view of the fusion rule of the traditional details injection method which causes larger spectral distortion, the characteristic product is chosen to express local image features for the high-frequency component obtained by the transform in this paper, and then the details of panchromatic image are extracted on the basis of the threshold of the decision-making factor, eventually effective high-frequency detail surface acquired by NSDFB inverse transform is attached to the component of multispectral image. The experimental results show that, the algorithm in this paper makes the spectral distortion minimum and makes correlation coefficient maxi- mum, and improves the spatial detail quality largely in a variety of different performance index evaluations.

关 键 词：图像融合 非下采样CONTOURLET变换 àtrous小波变换 特征量积 决策因子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]