基于FPDE的红外与可见光图像融合算法  

Fusion Algorithm of Infrared and Visible Images Based on FPDE

作　　者：高雪琴 刘刚 肖刚[2] BAVIRISETTI Durga Prasad 史凯磊 GAO Xue-Qin;LIU Gang;XIAO Gang;BAVIRISETTI Durga Prasad;SHI Kai-Lei(School of Automation Engineering,Shanghai University of Electrical Power,Shanghai 200090;School of Aeronau-tics and Astronautics,Shanghai Jiao Tong University,Shanghai 200240)

机构地区：[1]上海电力学院自动化工程学院,上海200090 [2]上海交通大学航空航天学院,上海200240

出　　处：《自动化学报》2020年第4期796-804,共9页Acta Automatica Sinica

基　　金：国家重点基础研究发展计划(973计划)(2014CB744903);国家自然科学基金(61673270,61203224);上海浦江人才计划(16PJD028)资助。

摘　　要：针对传统红外与可见光图像融合算法中存在的细节信息不够丰富,边缘信息保留不够充分等问题,文中提出了一种基于四阶偏微分方程(Fourth-order partial differential equation, FPDE)的改进的图像融合算法.算法首先采用FPDE将已配准的红外与可见光图像进行分解,得到高频分量和低频分量;然后,对高频分量采用基于主成分分析(Principal component analysis, PCA)的融合规则来得到细节图像,对低频分量采用基于期望值最大(Expectation maximization, EM)的融合规则来得到近似图像;最后,通过组合最终的高频分量和低频分量来重构得到最终的融合结果.实验是建立在标准的融合数据集上进行的,并与传统的和最近的融合方法进行比较,结果证明所提方法得到的融合图像比现有的融合方法能有效地综合红外与可见光图像中的重要信息,有更好的视觉效果.Aiming at the problems that the traditional infrared and visible light image fusion algorithms are not rich enough in detail,and the edge information is not sufficiently reserved, an improved image fusion algorithm based on the fourth-order partial differential equation(FPDE) is proposed. The algorithm firstly decomposes the registered infrared and visible images with FPDE to get the high frequency and low frequency components, and then uses the principal component analysis(PCA) fusion rules for the high frequency components to obtain the detail image. An approximation image is obtained using the fusion rule based on the expectation maximization(EM) for low frequency components;Finally,by combining the final high-frequency components and low frequency components to reconstruct the final fusion results.The experiment is based on a standard fusion dataset and compared with the traditional and recent fusion methods. The results show that the proposed fusion method can effectively integrate the important information of infrared and visible images, better visual effects.

关 键 词：图像融合 四阶偏微分方程 期望值最大 主成分分析 红外与可见光图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN219[自动化与计算机技术—计算机科学与技术]