基于全局稀疏梯度滤波器分解的图像融合方法  

Image fusion method based on global sparse gradient filter decomposition

作　　者：常莉红 冯福存[1] 罗徽 杨媛 陆万顺[1] CHANG Lihong;FENG Fucun;LUO Hui;YANG Yuan;LU Wanshun(School of Mathematics and Computer Science,Ningxia Normal University,Guyuan Ningxia 756099;School of Physics and Electronic Information Engineering,Ningxia Normal University,Guyuan Ningxia 756099)

机构地区：[1]宁夏师范学院数学与计算机科学学院,宁夏固原756099 [2]宁夏师范学院物理与电气工程学院,宁夏固原756099

出　　处：《宁夏师范学院学报》2022年第10期43-51,共9页Journal of Ningxia Normal University

基　　金：宁夏自然科学基金研究项目资助(2021AAC03235,2022AAC03334).

摘　　要：利用LIME的弱光图像增强方法和基于全局稀疏梯度的滤波器分解工具提出了一种弱光图像与红外线图像的融合方法,来提高融合结果中弱光图像的场景信息并保留红外线图像的高亮目标.首先利用LIME的图像增强方法对低照度图像进行增强;然后利用基于全局稀疏的梯度滤波器对增强的弱光图像和红外线图像进行分解得到了图像的近似层和剩余层.在近似层和剩余层分别选取局部能量和全局稀疏的梯度信息作为活动水平级进行融合;最后,将融合的近似层和剩余层叠加得到融合结果.实验结果表明所提算法对弱光图像的融合是十分有效的.A fusion method of low illumination image and infrared image is proposed by using the weak light image enhancement method based on LIME and decomposition tool based on the global sparse gradient filter to improve the scene information of low illumination image in the fusion result and retain the bright target of infrared image.Firstly, the low illumination image is enhanced using LIME method;Then the enhanced low illumination image and infrared image are decomposed into the approximation layer and residual layer of the image using the global sparse gradient filter.In the approximation layer and the residual layer, local energy and global sparse gradient information are selected as the activity level for fusion;Finally, the fusion results are obtained by superimposing the fused approximate layers and the remaining layers.The experiment results show that the proposed algorithm is very effective for the fusion of the low illumination image and infrared images.

关 键 词：图像融合 全局稀疏梯度滤波器 红外线图像 可见光图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]