摘要
提出了一种基于提升小波变换的图像融合新算法.该算法对多源图像进行提升小波分解,并针对分解后的低频和高频分量各自的特点,选用不同的规则进行融合,即低频近似系数采用基于邻域空间频率加一致性检验,高频细节系数采用基于像素点的绝对值取最大,最后进行提升小波逆变换得到融合图像.实验结果表明,当采用空间频率误差比率和平均梯度作为客观评价标准,该算法具有良好的融合效果.
In this paper, a new image fusion algorithm based on lifting wavelet transform is presented. The source images are decomposed using lifting wavelet transform respectively. Aiming at the coefficients of low frequency and high frequency, this algorithm choose a different rule to fuse the image. To the low frequency, the spatial frequency based on the neighborhood add consistency check is elected as the fusion guide. And the absolute maximum based on detail coefficients is selected as the guide to the high frequency. After that the fused image is obtained by using inverse lifting wavelet trans- form. Taking the ratio space frequency error and the mean gradient as criterions, experimental results demonstrate that the algorithm is very effective and is able to fuse multi-source images.
出处
《微电子学与计算机》
CSCD
北大核心
2009年第4期64-66,69,共4页
Microelectronics & Computer
基金
重庆市科委自然科学基金重点项目(8970)
关键词
图像融合
提升小波变换
空间频率
image fusion
lifting wavelet transform
space frequency
