摘要
提出了一种改进的拉普拉斯金字塔图像融合方法。首先将源图像进行高斯金字塔分解,接着再进行拉普拉斯金字塔分解,然后对由拉普拉斯金字塔形成的图像序列的对应层分别使用不同的融合规则进行融合。顶层采用加权融合,其余层采用基于区域能量的融合规则进行融合,最后再进行拉普拉斯反变换得到重构图像。试验结果表明,该方法能够有效的保留图像信息,融合效果良好。
An improved Laplacian pyramid image fusion method is proposed.First,the source image is decomposed by Gaussian pyramid,then,it is decomposed by Laplacian pyramid,and then the corresponding layers of image sequence formed by Laplacian pyramid are fused by using different fusion rules.The top layer is fused by using weighted fusion method,the rest of the layers are fused by using the rules based on regional energy,finally,the reconstructed image is obtained by the Laplacetransformation.Experiment result shows that this method can effectively keep the image information,and the fusion effect is good.
出处
《长江大学学报(自科版)(上旬)》
2016年第12期21-26,共6页
JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI & ENG
基金
国家自然科学基金项目(60572048)
湖北省自然科学基金重点项目(2013CFA053)