基于四阶微分方程的红外线-可见光图像融合方法  

Fusion Method Based on Fourth Order Differential for Visible-Infrared Image

作　　者：常莉红 马红芳 冯福存[1] 罗徽 CHANG Lihong;MA Hongfang;FENG Fucun;LUO Hui(School of Mathematics and Computer Science,Ningxia Normal University,Guyuan,Ningxia 756000;School of Physics and Electronic Information Engineering,Ningxia Normal University,Guyuan,Ningxia 756000)

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756000 [2]宁夏师范学院物理与电子信息工程学院,宁夏固原756000

出　　处：《绵阳师范学院学报》2024年第11期75-83,共9页Journal of Mianyang Teachers＇ College

基　　金：宁夏自然科学基金项目(2023AAC03330);宁夏高校科研项目(NYG2024182)。

摘　　要：利用光照估计方法增强低照度图像的场景信息进而提高与红外线图像的融合效果.采用图像增强技术和四阶微分方程的图像分解方法提出了一种图像融合方法.首先,对低照度图像利用基于光照估计的内容增强方法提高图像的场景细节.其次,对增强的可见光图像和红外线图像利用四阶微分方程进行分解得到近似部分和细节部分,在近似部分选取能量保护,细节部分采用梯度显著性图的融合规则进行融合;最后,利用简单的叠加得到最终的融合结果.将图像增强方法与四阶微分方程结合起来,通过不同的融合规则提高了融合结果,并通过实验证明所提方法能有效的保护图像的细节信息,提高融合效果.In order to further the fusion effect of visible and infrared images,an image fusion method was established by using fourth order differential equation and image enhancement.Firstly,the content enhancement method based on illumination estimation was used to improve the scene details of low illumination images.Sec-ondly,the enhanced visible image and infrared image were decomposed by the fourth-order differential equation using the gradient descent method to get the approximate part and the detail part.The energy protection and the gra-dient saliency map rules were used in the approximate part and the detail part;Finally,the final fusion result was obtained by simple superposition.Experimental results showed that the proposed method could effectively protect the details of the image and improve the fusion effect.

关 键 词：图像融合 四阶微分方程 光照估计 图像增强 

分 类 号：TN911.7[电子电信—通信与信息系统]