摘要
将图像的高频信息和低频信息分开考虑,它们有各自的融合规则.首先通过离散小波变换获得高频系数部分,非下采样轮廓波变换获得低频和高频信息进行初步融合,最后与离散小波变换的结果进行二次融合.实验结果表明,与几种优秀算法比较,此算法表现出更好的视觉融合效果和更短的运行时间.
In this paper,the high-frequency and low-frequency information were considered separately,and they both have their own fusion rules.First we use discrete wavelet transform to get high frequency coefficients part,and we use non-subsampled contourlet transform to obtain low and high frequency parts to get an initial fusion image,then a second fusion with the result of discrete wavelet transform.Experimental results show that compared with the several excellent algorithms,this algorithm shows better visual effects and shorter running time.
出处
《微电子学与计算机》
CSCD
北大核心
2015年第9期40-44,共5页
Microelectronics & Computer
基金
河南省科技厅软科学研究计划项目(142400411213)
河南省高等学校重点科研项目(15A520118)
河南省信息技术教育研究项目(ITE2172)
关键词
图像融合
离散小波变换
非下采样轮廓波变换
二次融合
image fusion
discrete wavelet transform
non-subsampled coutourlet transform
second fusion
