摘要
针对多聚焦图像融合,在小波变换基础上提出了一种多策略的融合方法。图像小波变换后的低频分量采用基于清晰度的方法,清晰度评价函数选用八邻域拉普拉斯算子和;而高频分量采用基于空间频率的方法,并且只计算与高频方向相一致的空间频率,减小了计算量。融合效果评价除使用传统方法的熵、交叉熵外,还使用了基于三个方向(水平、垂直和对角)的空间频率和通用的主观与客观相结合的新方法。实验结果表明,本文方法得到的融合图像效果最好,优于文中的其它算法,同时本文使用的评价方法也是适用和有效的。
A method of multi-decision on the basis of wavelet transform was presented for multi-focus image fusion, After wavelet transform of image, the definition was applied in low frequency part, and the spatial frequency was applied in high frequency parts. The evaluating function of definition was the sum of Laplacian operator in the eight directions. The spatial frequency was calculated only in one direction that was the same as high frequency direction, and this method can decrease the workload, Besides the traditional evaluating methods of fused image quality, such as entropy and cross entropy, the spatial frequency was applied based on three directions which were horizontal, vertical and diagonal, and the new universal method which combined objective and subjective factors, was also used. The experimental results show that the fused image obtained by proposed method has best effect, and the proposed fusion scheme is superior to the other algorithms mentioned in this paper.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2007年第8期93-98,共6页
Opto-Electronic Engineering
基金
国家自然科学基金资助项目(60475036)
关键词
图像融合
小波变换
空间频率
图像清晰度
image fusion
wavelet transform
spatial frequency
image definition
