摘要
以提升遥感图像和多聚焦图像的融合精度为目的,结合非下采样剪切波变换(NSST)可以捕捉图像的细节特征,提出了一种NSST和加权区域特性的图像融合方法。利用非下采样剪切波变换对源图像进行多尺度、多方向分解,得到低频子带和高频子带,低频子带系数采用改进梯度投影的非负矩阵分解(NMF),高频子带系数采用加权区域能量和区域方差相结合的融合策略,然后应用非下采样剪切波的逆变换得到融合的图像。实验结果表明:该方法从主观视觉方面很好地保留了多幅图像的有用信息,给出该方法与其他融合算法在客观评价指标应用信息熵EN、互信息MI和加权边缘信息量QAB/F的比较结果 。
To improve the remote sensing image and multi-focus image fusion accuracy,combined with the non-sampling shearlets transform(NSST)which can capture the details of the image features,we proposed a image fusion method based on NSST and weighted area feature.Firstly this method uses non-sampling shearlets transform for source image to carry on multiscale multi-direction decomposition to get low-frequency and high-frequency sub-bands.Then the improved gradient projection of non-negative matrix factorization(NMF)is used for the low-frequency sub-band coefficient,while the high frequency sub-band coefficient uses the fusion strategy combining the regional energy and variance of the weighted area.Finally,the non-sampling shearlets inverse transformation is used to get image fusion.The experimental results show that this method can well retain the useful information of multiple images from the aspect of subjective vision,and the comparison results are given with other fusion methods from the aspect of objective evaluation indexes such as entropy,mutual information and weighted edge information preservation values.
出处
《应用光学》
CAS
CSCD
北大核心
2015年第5期735-741,共7页
Journal of Applied Optics
基金
陕西省自然科学基金(2014JQ1008)
关键词
非抽样剪切波变换
非负矩阵分解
融合策略
加权区域能量
non-sampling shearlets
non-negative matrix decomposition
fusion strategy
weighted area energy
