摘要
针对遥感图像融合结果光谱信息丰富度欠缺的问题,提出了一种基于结构提取的融合方法。首先,采用一种通过相对总变差从纹理中提取结构的方法对图像进行分解,分离出对应的低频和高频子带。其次,在高频层引入视觉决策图来指导高频进行融合,在低频融合规则中,同时考虑全色图像的边缘强度和多光谱的色彩信息强度来进行融合。最后,通过重构低频融合结果和高频融合结果得到融合图像。在Worldview-2卫星遥感图像上验证本文算法的有效性。实验表明:本文算法较好地保留了全色图像的细节信息以及多光谱图像的光谱信息,与其他同类算法相比具有一定的优越性。
Aiming at taking both color information and edge retention into count as much as possible,an effective multi-scale fusion method based on structure extraction and visual salience map(VSM)is proposed.Firstly,the source images are decomposed by a validated structure extraction from the texture via relative total variation.Then,preserving the color of MS and the edge definition are considered in the fusion of base layers and the conventional'averaging'fusion scheme is unable to achieve the same effects.Next,VSM-based technique is proposed to fuse the detail layers to figure out perceptually and combine salient visual structures,standing out from neighbors.Finally,the desired fusion image was obtained by image reconstruction,using the detailed and based fused results.Remote sensing images from the Worldview-2 satellite are used as the data set which verifies the effectiveness of our algorithm.Experiments show that the algorithm designed in this paper retains the detail information of the panchromatic image and the spectral information of multispectral image,and highlight the expediency and suitability with comparing with others.
作者
李雄飞
吴佳婧
张小利
王泽宇
冯云丛
LI Xiong-fei;WU Jia-jing;ZHANG Xiao-li;WANG Ze-yu;FENG Yun-cong(College of Software,Jilin University,Changchun 130012,China;Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education,Jilin University,Changchun 130012,China;College of Computer Science and Engineering,Changchun University of Technology,Changchun 130012,China)
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2021年第5期1775-1784,共10页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(61801190)
吉林省自然科学基金项目(20180101055JC)
国家博士后科研基金项目(2017M611323)
吉林省“十三五”教育科研规划项目(JJKH2020997KJ,JJKH20200678KJ)
中央高校基本科研基金项目(93K172020K05).
关键词
图像融合
遥感图像
结构提取
融合规则
image fusion
remote sensing images
structure extraction
fusion rules
