基于FPDE的红外与可见光图像融合算法

高雪琴; 刘刚; 肖刚; BAVIRISETTIDurga Prasad; 史凯磊

doi:10.16383/j.aas.2018.c180188

基于FPDE的红外与可见光图像融合算法

doi: 10.16383/j.aas.2018.c180188

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

基金项目:

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

国家自然科学基金 61203224

上海浦江人才计划 16PJD028

详细信息

作者简介:
高雪琴上海电力学院自动化工程学院硕士研究生.主要研究方向为图像融合. E-mail: gxq0163@163.com

肖刚上海交通大学航空航天学院教授. 2004年获得上海交通大学博士学位.主要研究方向为先进航空电子综合仿真, 智能信息处理.E-mail:xiaogang@sjtu.edu.cn

BAVIRISETTIDurga Prasad：BAVIRISETTI Durga Prasad 上海交通大学航空航天学院博士后. 2016年获得印度韦洛尔科技大学博士学位.主要研究方向为图像融合, 目标跟踪, 机器学习和计算机视觉. E-mail:bdps1989@sjtu.edu.cn

史凯磊上海电力学院自动化工程学院硕士研究生.主要研究方向为机器视觉. E-mail: 18321779651@163.com

通讯作者:
刘刚上海电力学院自动化工程学院教授. 2005年获得上海交通大学博士学位.主要研究方向为模式识别, 机器学习.本文通信作者. E-mail: liugang@shiep.edu.cn

计量
- 文章访问数: 3374
- HTML全文浏览量: 2204
- PDF下载量: 310
- 被引次数: 0
出版历程
- 收稿日期: 2018-04-02
- 录用日期: 2018-09-12
- 刊出日期: 2020-04-24

Fusion Algorithm of Infrared and Visible Images Based on FPDE

1.
School of Automation Engineering, Shanghai University of Electrical Power, Shanghai 200090
2.
School of Aeronautics and Astronautics, Shanghai Jiao Tong University, Shanghai 200240

Funds:

National Basic Research Program of China (973 Program) 2014CB744903

National Natural Science Foundation of China 61673270

National Natural Science Foundation of China 61203224

Shanghai Pujiang Talent Program 16PJD028

More Information

Author Bio:
GAO Xue-Qin Master student at the School of Automation Engineering, Shanghai University of Electrical Power, China. Her main research interest is image fusion

XIAO Gang Professor at the School of Aeronautics and Astronautics, Shanghai Jiao Tong University, China. He received his Ph. D. degree from Shanghai Jiao Tong University, China in 2004. His research interest covers advanced avionics integrated simulation and intelligent information processing

BAVIRISETTI Durga Prasad Postdoctor at the School of Aeronautics and Astronautics, Shanghai Jiao Tong University. He received his Ph. D. degree from Vellore Institute of Technology, India. His research interest covers image fusion, target tracking, machine learning, and computer vision

SHI Kai-Lei Master student at the School of Automation Engineering, Shanghai University of Electrical Power, China. His main research interest is machine vision

Corresponding author: LIU Gang Professor at the School of Automation Engineering, Shanghai University of Electrical Power, China. He received his Ph. D. degree from Shanghai Jiao Tong University, China in 2005. His research interest covers pattern recognition and machine learning. Corresponding author of this paper

摘要

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.
- 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
Abstract: 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.
- Image fusion /
- fourth-order partial differential equation(FPDE) /
- expectation maximization (EM) /
- principal component analysis(PCA) /
- infrared and visible images
Recommended by Associate Editor HUANG Qing-Ming
注释:

1) 本文责任编委黄庆明

HTML全文

图 1 本文算法的流程图

Fig. 1 Flow chart of the algorithm of this paper

下载: 全尺寸图片幻灯片

图 2 "dune"图像的融合结果

Fig. 2 The fusion result of image "dune"

下载: 全尺寸图片幻灯片

图 3 "pavilion"图像的融合结果

Fig. 3 The fusion result of image "pavilion"

下载: 全尺寸图片幻灯片

图 4 "maninhuis"图像的融合结果

Fig. 4 The fusion result of image "maninhuis"

下载: 全尺寸图片幻灯片

图 5 "UN Camp"图像的融合结果

Fig. 5 The fusion result of image "UN Camp"

下载: 全尺寸图片幻灯片

图 6 其他6组图像的融合结果

Fig. 6 The fusion result of other six groups of images

下载: 全尺寸图片幻灯片

表 1 "dune"的融合结果的客观指标评价结果

Table 1 Objective evaluation result of the fusion result of "dune"

融合方法	MI	IE	SD	EIPV
GRAD	1.2948	4.9	37.9309	0.4278
RATIO	0.7248	4.2485	18.5883	0.3546
DWT	0.6542	4.3821	21.0136	0.3841
FPDE	0.7339	4.2997	19.6550	0.4821
CVT-SR	1.5809	4.8953	38.4630	0.4408
VSM	0.7053	4.3278	20.1876	0.4085
本文方法	2.4647	5.0006	41.2892	0.5106

下载: 导出CSV

表 2 "pavilion"的融合结果的客观指标评价结果

Table 2 Objective evaluation result of the fusion result of "pavilion"

融合方法	MI	IE	SD	EIPV
GRAD	1.9984	4.4074	26.9211	0.4625
RATIO	1.2350	4.4073	25.5996	0.3724
DWT	1.1351	4.4671	28.2542	0.4197
FPDE	1.3621	4.3829	25.8059	0.3846
CVT-SR	1.3812	5.2110	48.8611	0.4251
VSM	1.2409	4.4436	27.8868	0.4451
本文方法	3.1240	4.8207	59.5605	0.4671

下载: 导出CSV

表 3 "maninhuis"的融合结果的客观指标评价结果

Table 3 Objective evaluation result of the fusion result of "maninhuis"

融合方法	MI	IE	SD	EIPV
GRAD	1.6169	4.7116	28.0854	0.4814
RATIO	0.9408	4.6744	26.7289	0.3857
DWT	0.9312	4.7941	30.4888	0.4371
FPDE	1.0689	4.7081	27.4131	0.4516
CVT-SR	1.3302	5.151	46.7314	0.5015
VSM	1.0296	4.760	29.3688	0.4921
本文方法	3.2612	5.036	57.1928	0.5091

下载: 导出CSV

表 4 "UN Camp"的融合结果的客观指标评价结果

Table 4 Objective evaluation result of the fusion result of "UN Camp"

融合方法	MI	IE	SD	EIPV
GRAD	1.3059	4.8571	36.0841	0.3918
RATIO	1.0455	4.3359	22.9454	0.3895
DWT	1.0314	4.4535	24.9950	0.3641
FPDE	1.0894	4.3503	22.9777	0.4234
CVT-SR	1.1538	4.7549	32.5279	0.3635
VSM	1.0753	4.3991	23.9920	0.4064
本文方法	2.5529	4.9069	41.5711	0.4461

下载: 导出CSV

表 5 其他6组图像客观评价结果

Table 5 Objective evaluation result of the other six groups of images

融合方法	GRAD	RATIO	DWT	FPDE	CVT-SR	VSM	本文方法
MI	1.9322	1.3495	1.2301	1.3235	1.4803	1.3201	2.9112
IE	4.9659	4.3952	4.4904	4.4409	4.8827	4.4594	4.9711
SD	49.5321	23.9142	25.5283	24.3822	41.2654	26.6849	51.8051
EIPV	0.4732	0.3491	0.4651	0.5244	0.4653	0.4857	0.5276

下载: 导出CSV

表 6 各种融合方法的计算时间对比

Table 6 Computational time comparison of different fusion methods

融合方法	GRAD	RATIO	DWT	FPDE	CVT-SR	VSM	本文方法
时间(s)	2.7914	0.6909	1.6567	8.1072	5.4289	5.6316	71.7869

下载: 导出CSV

参考文献(22)

[1]	童涛, 杨桄, 孟强强, 孙嘉成, 叶怡, 陈晓榕.基于边缘特征的多传感器图像融合算法.红外与激光工程, 2014, 43(1): 311-317 doi: 10.3969/j.issn.1007-2276.2014.01.055 Tong Tao, Yang Guang, Meng Qiang-Qiang, Sun Jia-Cheng, Ye Yi, Chen Xiao-Rong. Multi-sensor image fusion algorithm based on edge feature. Infrared and Laser Engineering, 2014, 43(1): 311-317 doi: 10.3969/j.issn.1007-2276.2014.01.055
[2]	Li Shu-Tao, Kang Xu-Dong, Fang Le-Yuan, Hu Jian-Wen, Yin Hai-Tao. Pixel-level image fusion: A survey of the state of the art. Information Fusion, 2017, 33(1): 100-112
[3]	Zhang Qiong, Maldague Xavier. An adaptive fusion approach for infrared and visible images based on NSCT and compressed sensing. Infrared Physics and Technology, 2016, 74(1): 11-20 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=0b634893125f83498d0b8613c8c9ec69
[4]	Lee M W, Kwak K C. Performance Comparison of Infrared and Visible Image Fusion Approaches. In: Proceedings of the 2017 International Conference on Control, Artificial Intelligence, Robotics & Optimization (ICCAIRO), Prague, Czech Republic: 2017. 274-277
[5]	Burt P J, Adelson E H.The Laplacian pyramid as a compact image code. IEEE Transactions on Communications, 1983, 31(4): 532-540 doi: 10.1109/TCOM.1983.1095851
[6]	肖进胜, 饶天宇, 贾茜, 宋金钟, 易本顺.基于图切割的拉普拉斯金字塔图像融合算法.光电子·激光, 2014, 25(07): 1416-1424 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gdzjg201407032 Xiao Jin-Sheng, Rao Tian-Yu, Jia Qian, Song Jin-Zhong, Yi Ben-Shun. An image fusion algorithm of Laplacian pyramid based on graph cuting. Journal of Optoelectronics·Laser, 2014, 25(07): 1416-1424 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=gdzjg201407032
[7]	Burt P J, A gradient pyramid basis for pattern-selective image fusion, In: Proceedings of the SID International Symposium. New York, USA: 1992. 467-470
[8]	Alexander Toet, Image fusion by a ratio of low-pass pyramid. Pattern Recognition Letters, 1989, 9(4): 245-253 doi: 10.1016/0167-8655(89)90003-2
[9]	李建林, 俞建成, 孙胜利.基于梯度金字塔图像融合的研究.科学技术与工程, 2007, 7(22): 5818-5822 doi: 10.3969/j.issn.1671-1815.2007.22.018 Li Jian-Lin, Yu Jian-Cheng, Sun Sheng-Li. Research on image fusion based on gradient pyramid. Science Technology and Engineering, 2007, 7(22): 5818-5822 doi: 10.3969/j.issn.1671-1815.2007.22.018
[10]	Xiang Yan, Qin Han-Lin, Li Jia, Zhou Hui-Xin, Zong Jing-Guo. Infrared and visible image fusion with spectral graph wavelet transform. Journal of the Optical Society of America A Optics Image Science & Vision, 2015, 32(9): 1643-1652 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=0da6606f5b60ddc886110a710f7e7b12
[11]	Pal C, Das P, Chakrabarti A, Ghosh R. Rician noise removal in magnitude MRI images using efficient anisotropic diffusion filtering. International Journal of Imaging Systems and Technology, 2017, 27(3): 248-264 doi: 10.1002/ima.22230
[12]	钱伟新, 刘瑞根, 王婉丽, 祁双喜, 王伟, 程晋明.基于图像特征方向的各向异性扩散滤波方法.中国图象图形学报, 2006, 11(6): 818-822 doi: 10.3969/j.issn.1006-8961.2006.06.008 Qian Wei-Xin, Liu Rui-Gen, Wang Wan-Li, Qi Shuang-Xi, Wang Wei, Cheng Jin-Ming. The anisotropic diffusion methods based on the directions of the image feature. Journal of Image and Graphics, 2006, 11(6): 818-822 doi: 10.3969/j.issn.1006-8961.2006.06.008
[13]	You Yu-Li, Kaveh M. Fourth-order partial differential equations for noise removal. IEEE Transactions on Image Processing, 2000, 9(10): 1723-1730 doi: 10.1109/83.869184
[14]	Bavirisetti D P, Xiao G, Liu G. Multi-sensor image fusion based on fourth order partial differential equations. International Conference on Information Fusion. Xi'an, China: IEEE, 2017. 1-9
[15]	Strang G. Introduction to Applied Mathematic, New York, Wellesley-Cambridge. Springer, 1986. 166-179
[16]	Himanshi, Bhateja V, Krishn A, Sahu A. An improved medical image fusion approach using PCA and complex wavelets. In: Proceedings of International Conference on Medical Imaging, M-Health and Emerging Communication Systems, Greater Noida, India: IEEE, 2015. 442-447
[17]	Yang J, Blum R S. A statistical signal processing approach to image fusion for concealed weapon detection. In: Proc- eedings of the 2002 International Conference on Image Processing, New York, USA: IEEE, 2002. 513-516
[18]	TNO Image Fusion Dataset[Online], available: https://figshare.com/articles/TNO_Image_Fusion_Dataset/1008029, September 15, 2018
[19]	Kadar I. Pixel-level image fusion: the case of image sequences. In: Proceedings of SPIE-The International Society for Optical Engineering, 1998, 3374: 378-388
[20]	Ma Jin-Lei, Zhou Zhi-Qiang, Wang Bo, Zong Hua. Infrared and visible image fusion based on visual saliency map and weighted least square optimization. Infrared Physics & Technology, 2017, 82: 8-17 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=0cf016ce28d21ceabc85edf6d555f29c
[21]	Liu Yu, Liu Shu-Ping, Wang Zeng-Fu. A general framework for image fusion based on multi-scale transform and sparse representation. Information Fusion, 2015, 24: 147-164 doi: 10.1016/j.inffus.2014.09.004
[22]	张小利, 李雄飞, 李军.融合图像质量评价指标的相关性分析及性能评估.自动化学报, 2014, 40(2): 306-315 doi: 10.3724/SP.J.1004.2014.00306 Zhang Xiao-Li, Li Xiong-Fei, Li Jun. Validation and correlation analysis of metrics for evaluating performance of image fusion. Acta Automatica Sinica, 2014, 40(2): 306-315 doi: 10.3724/SP.J.1004.2014.00306