光学精密工程2023，Vol.31Issue(23)：3490-3503,14.DOI:10.37188/OPE.20233123.3490
基于目标增强与蝴蝶优化的红外与可见光图像融合
Infrared and visible image fusion based on target enhancement and butterfly optimization
摘要
Abstract
A fusion method for infrared and visible images was developed based on target enhancement and butterfly optimization to address the problems of target blurring,textural detail loss,and artifacts in traditional infrared and visible image fusion results.First,to deal with the blurring of the fusion target edge caused by the imaging mechanism of the infrared image,an infrared image enhancement module based on target edge enhancement was constructed.Second,a visible light enhancement module based on multiscale Retinex with color restoration was developed to solve the problem of missing details in fused images caused by the low quality of visible images.Third,fourth-order partial differential equa-tions and the principal component analysis method were used to smooth out the edges of infrared and visi-ble enhanced images to solve the problem of artifacts in the fusion results.Finally,an image reconstruc-tion module based on butterfly optimization was designed for the adaptive allocation of reconstructed im-age weights.This allows the target to be highlighted in the final fusion result while retaining more textur-al details.To verify the advantages of the proposed algorithm,it was compared with six classic algo-rithms on the TNO,INO,M3FD,and RoadScene datasets.The experimental results show that the fu-sion results obtained by the proposed algorithm have clear edges,strong contrast,no artifacts,and rich textural details.Compared with the other algorithms,the objective evaluation indicators EN,SF,SD,JE,VIF,and NSS are improved by an average of 9.24%,38.88%,51.11%,4.65%,35.44%and 19.36%,respectively.关键词
图像融合/红外图像/可见光图像/多尺度/蝴蝶优化Key words
image fusion/infrared image/visible image/multi-scale/butterfly optimization分类
计算机与自动化引用本文复制引用
郝帅,李彤,马旭,何田,孙曦子,李嘉豪..基于目标增强与蝴蝶优化的红外与可见光图像融合[J].光学精密工程,2023,31(23):3490-3503,14.基金项目
国家自然科学基金资助项目(No.51804250) （No.51804250）
中国博士后科学基金资助项目(No.2019M653874XB,No.2020M683522) （No.2019M653874XB,No.2020M683522）
陕西省科技计划项目(No.2021JQ-572,No.2020JQ-757) （No.2021JQ-572,No.2020JQ-757）
陕西省教育厅科研计划项目(No.18JK0512) （No.18JK0512）
