朱浩然¹,
刘云清^1,,,
张文颖^{2, 3}
1.长春理工大学电子信息工程学院 ??长春 ??130022
2.长春理工大学光电工程学院 ??长春 ??130022
3.中国科学院光电研究院 ??北京 ??100094

详细信息

作者简介:朱浩然：男，1987年生，博士生，研究方向为图像融合、图像增强等
刘云清：男，1970年生，教授，博士生导师，主要研究方向为自动控制与测试技术等
张文颖：女，1988年生，博士生，研究方向为光电测量与精密仪器等

通讯作者:刘云清　mzliuyunqing@163.com

中图分类号:TP391

计量

文章访问数:1490
HTML全文浏览量:519
PDF下载量:63
被引次数:0

出版历程

收稿日期:2017-05-02
修回日期:2018-10-18
网络出版日期:2018-10-31
刊出日期:2019-03-01

Night-vision Image Fusion Based on Intensity Transformation and Two-scale Decomposition

Haoran ZHU¹,
Yunqing LIU^1,,,
Wenying ZHANG^{2, 3}
1. School of Electronics and Information Engineering, Changchun University of Science and Technology, Changchun 130022, China
2. Photoelectric Engineering College, Changchun University of Science and Technology, Changchun 130022, China
3. Academy of Opto-electronics, Chinese Academy of Sciences, Beijing 100094, China


摘要
摘要:为了获得更适合人感知的夜视融合图像，该文提出一种基于灰度变换与两尺度分解的夜视图像融合算法。首先，利用红外像素值作为指数因子对可见光图像进行灰度转换，在达到可见光图像增强的同时还使可见光与红外图像融合任务转换为同类图像融合。其次，通过均值滤波对增强结果与原始可见光图像进行两尺度分解。再次，运用基于视觉权重图的方法融合细节层。最后，综合这些结果重构出融合图像。由于该文方法在可见光波段显示结果，因此融合图像更适合视觉感知。实验结果表明，所提方法在视觉质量和客观评价方面优于其它5种对比方法，融合时间小于0.2 s，满足实时性要求。融合后图像背景细节信息清晰，热目标突出，同时降低处理时间。
关键词:图像融合/
灰度变换/
两尺度分解/
视觉显著性检测
Abstract:In order to achieve more suitable night vision fusion images for human perception, a novel night-vision image fusion algorithm is proposed based on intensity transformation and two-scale decomposition. Firstly, the pixel value from the infrared image is used as the exponential factor to achieve intensity transformation of the visible image, so that the task of infrared-visible image fusion can be transformed into the merging of homogeneous images. Secondly, the enhanced result and the original visible image are decomposed into base and detail layers through a simple average filter. Thirdly, the detail layers are fused by the visual weight maps. Finally, the fused image is reconstructed by synthesizing these results. The fused image is more suitable for the visual perception, because the proposed method presents the result in the visual spectrum band. Experimental results show that the proposed method outperforms obviously the other five methods. In addition, the computation time of the proposed method is less than 0.2 s, which meet the real-time requirements. In the fused result, the details of the background are clear while the objects with high temperature variance are highlighted as well.
Key words:Image fusion/
Intensity transformation/
Two-scale decomposition/
Visual saliency detection



PDF全文下载地址:

https://jeit.ac.cn/article/exportPdf?id=d6ac0b41-c178-4916-bc92-9a7745541cb0