一种基于ResNet152的红外与可见光图像融合算法 
下载： 1389次
An Infrared and Visible Image Fusion Algorithm Based on ResNet152
1 兰州交通大学电子与信息工程学院, 甘肃 兰州 730070
2 兰州交通大学测绘与地理信息学院, 甘肃 兰州 730070
3 地理国情监测技术应用国家地方联合工程研究中心, 甘肃 兰州 730070
图 & 表
图 1. 残差块结构
Fig. 1. Structure of residual block
下载图片 查看原文
图 2. 本文方法的框架图
Fig. 2. Framework of proposed method
下载图片 查看原文
图 3. 高频部分融合流程图
Fig. 3. Flowchart of high frequency part fusion
下载图片 查看原文
图 4. 卷积运算图
Fig. 4. Convolution operation
下载图片 查看原文
图 5. 不同融合方法的比较
Fig. 5. Comparison of different fusion methods
下载图片 查看原文
表 1图像融合的客观评价
Table1. Image fusion objective evaluation
| Image | Metrics | CBF | WLS | JSR | JSRSD | ConvSR | Proposed |
|---|
| a | EFMIpixelEFMIdctEFMIwNabfESSIM | 0.87010.24500.29840.23171.2476 | 0.88470.27450.33510.16211.4442 | 0.85280.15380.21130.23321.2146 | 0.83390.13150.18120.32851.0792 | 0.83360.14220.32050.12241.1968 | 0.88560.33780.38280.02501.5324 | | b | EFMIpixelEFMIdctEFMIwNabfESSIM | 0.89680.22290.29200.25541.2994 | 0.93050.28900.34950.20911.4618 | 0.91890.16860.23260.24301.2562 | 0.89320.14430.20360.34191.1330 | 0.91600.13880.38130.05721.3608 | 0.93450.35630.39060.03741.5454 | | c | EFMIpixelEFMIdctEFMIwNabfESSIM | 0.86450.22990.26330.36071.0740 | 0.89180.29810.34640.21561.3834 | 0.87920.16990.22050.17691.2560 | 0.86320.14540.19560.24171.1874 | 0.84770.13420.34240.08681.2084 | 0.89270.34770.38360.03871.4620 | | d | EFMIpixelEFMIdctEFMIwNabfESSIM | 0.83860.18350.24070.52780.9150 | 0.89770.32360.37730.28211.3538 | 0.88850.15910.22660.21091.2506 | 0.86530.13250.20780.28011.1616 | 0.83800.10650.29920.11451.1500 | 0.91280.36490.40070.03761.4478 | | e | EFMIpixelEFMIdctEFMIwNabfESSIM | 0.85690.37810.46500.15231.0472 | 0.84710.39080.44280.28201.0916 | 0.84810.23600.28400.18730.9764 | 0.84340.20990.25650.21980.9374 | 0.86290.30610.51440.03481.1292 | 0.86340.44290.44760.04831.1808 |
|
查看原文
李恒, 张黎明, 蒋美容, 李玉龙. 一种基于ResNet152的红外与可见光图像融合算法[J]. 激光与光电子学进展, 2020, 57(8): 081013. Heng Li, Liming Zhang, Meirong Jiang, Yulong Li. An Infrared and Visible Image Fusion Algorithm Based on ResNet152[J]. Laser & Optoelectronics Progress, 2020, 57(8): 081013.
PDF全文
