红外技术2026，Vol.48Issue(4)：438-448,11.
基于快速梯度域引导滤波的红外与可见光图像融合
Fusion of Infrared and Visible Images Based on Fast Gradient Domain Guided Filtering
摘要
Abstract
To address the issue of edge information loss in infrared and visible light image fusion under traditional multi-scale domain fusion rules,this study proposes an infrared and visible light image fusion method based on a Fast Gradient Domain Guided Filter(FGDGF).Building upon the Gradient Domain Guided Filter(GDGF),the FGDGF method adjusts the filter scale to preserve the structural information and denoising capabilities of the image while effectively maintaining large-gradient edge information,thereby achieving superior edge preservation performance and improved computational efficiency.First,the input source images are decomposed using a fast gradient domain guided filtering approach.Subsequently,a weighted fusion rule based on Visual Saliency Mapping(VSM)is applied to generate the base layer fusion image.For the detail layer,an adaptive Pulse Coupled Neural Network(PCNN)fusion rule with optimized parameters is employed.Finally,the final fusion image is reconstructed by combining the base and detail layers.Experimental validation indicates that the proposed method achieves notable improvements in objective evaluation metrics.Specifically,the average gradient,correlation coefficient,information entropy,spatial frequency,and standard deviation are enhanced by 28.6%,14.9%,8.9%,32.6%,and 11.4%on average,respectively.These results indicate that the method not only effectively preserves edge and texture information from the source images but also enhances visual quality and operational efficiency.关键词
图像处理/快速梯度域引导滤波/红外与可见光图像融合/脉冲耦合神经网络Key words
image processing/fast gradient domain guided filter/infrared and visible image fusion/pulse coupled neural network分类
信息技术与安全科学引用本文复制引用
杨艳春,王泽煜,李毅,李佳龙..基于快速梯度域引导滤波的红外与可见光图像融合[J].红外技术,2026,48(4):438-448,11.基金项目
长江学者和创新团队发展计划资助(IRT_16R36),国家自然科学基金(62462043),甘肃省科技计划项目(18JR3RA104),甘肃省高等学校产业支撑计划项目(2020C-19),甘肃省重点研发计划(NO.25YFGA047). （IRT_16R36）
