• 论文 • 上一篇    下一篇

基于Huber正则化的红外与可见光图像融合

杨文莉1, 黄忠亿2   

  1. 1. 中国矿业大学数学学院, 徐州 221116;
    2. 清华大学数学科学系, 北京 100084
  • 收稿日期:2021-09-14 出版日期:2022-07-14 发布日期:2022-08-03
  • 基金资助:
    国家自然科学基金(12001529,12025104,11871298,81930119)资助.

杨文莉, 黄忠亿. 基于Huber正则化的红外与可见光图像融合[J]. 计算数学, 2022, 44(3): 305-323.

Yang Wenli, Huang Zhongyi. INFRARED AND VISIBLE IMAGE FUSION BASED ON HUBER REGULARIZATION[J]. Mathematica Numerica Sinica, 2022, 44(3): 305-323.

INFRARED AND VISIBLE IMAGE FUSION BASED ON HUBER REGULARIZATION

Yang Wenli1, Huang Zhongyi2   

  1. 1. School of Mathematics, China University of Mining and Technology, Xuzhou 221116, China;
    2. Department of Mathematics, Tsinghua University, Beijing 100084, China
  • Received:2021-09-14 Online:2022-07-14 Published:2022-08-03
图像融合通常是指从多源信道采集同一目标图像,将互补的多焦点、多模态、多时相和/或多视点图像集成在一起,形成新图像的过程.在本文中,我们采用基于Huber正则化的红外与可见光图像的融合模型.该模型通过约束融合图像与红外图像相似的像素强度保持热辐射信息,以及约束融合图像与可见光图像相似的灰度梯度和像素强度保持图像的边缘和纹理等外观信息,同时能够改善图像灰度梯度相对较小区域的阶梯效应.为了最小化这种变分模型,我们结合增广拉格朗日方法(ALM)和量身定做有限点方法(TFPM)的思想设计数值算法,并给出了算法的收敛性分析.最后,我们将所提模型和算法与其他七种图像融合方法进行定性和定量的比较,分析了本文所提模型的特点和所提数值算法的有效性.
Image fusion usually refers to the process of acquiring the same scene from multiple source channels and integrating complementary multi-focus, multi-modal, multi-temporal and/or multi-viewpoint images into a new image. In this paper, we propose using the Huber regularization based infrared and visible image fusion model. It maintains thermal radiation information by constraining the fused image to have similar pixel intensity with the infrared image, and keeps the appearance information such as the edges and texture of the source images by constraining the fused image to have similar gray gradient and pixel intensity with the visible image, and it can also ameliorate the staircase for the areas of the fused images with a relatively small gray gradient. To minimize the proposed model, we combined the augmented Lagrangian method (ALM) and the tailored finite point method (TFPM) to design the numerical algorithm, and we establish the convergence analysis of the proposed algorithm. Numerical experiments are compared qualitatively and quantitatively with other seven image fusion methods, to demonstrate the features of our model and show the efficiency of the proposed numerical method.

MR(2010)主题分类: 

()
[1] Meyer Y. Oscillating patterns in image processing and nonlinear evolution equations:the fifteenth Dean Jacqueline B. Lewis memorial lectures[M]. American Mathematical Soc., 2001.
[2] Bellettini G, Caselles V, Novaga M. The total variation flow in RN[J]. J. Differ. Equations, 2002, 184(2):475-525.
[3] Rudin L I, Osher S, Fatemi E. Nonlinear total variation based noise removal algorithms[J]. Physica D., 1992, 60:259-268.
[4] Ma Y, Chen J, Chen Ch, Fan F, Ma J Y. Infrared and visible image fusion using total variation model[J]. Neurocomputing, 2016, 202:12-19.
[5] Wu C, Tai X C. Augmented Lagrangian method, dual methods, and split Bregman iteration for ROF, vectorial TV, and high order models[J]. SIAM J. Imaging Sci., 2010, 3(3):300-339.
[6] Boyd S, Parikh N, Chu E. Distributed optimization and statistical learning via the alternating direction method of multipliers[J]. Found. Trends Mach. le., 2011, 3(1):1-122.
[7] Zhu W. A first-order image denoising model for staircase reduction[J] Adv. Comput. Math., 2019, 45:3217-3239.
[8] Ma J, Chen C, Li C, Huang J. Infrared and visible image fusion via gradient transfer and total variation minimization[J]. Inform. Fusion, 2016, 31:100-109.
[9] Ma Y, Chen J, Chen C, Fan F, Ma J. Infrared and visible image fusion using total variation model[J]. Neurocomputing, 2016, 202:12-19.
[10] Da Cunha A L, Zhou J, Do M N. The Nonsubsampled Contourlet Transform:Theory, Design, and Applications[J]. IEEE T. Image Process., 2006, 15(10):3089-3101.
[11] Da Cunha A L, Zhou J, Do M N. Nonsubsampled contourilet transform:filter design and applications in denoising[J]. Proc. IEEE Int. Conf. Image Process., 2005, 1:1-749.
[12] Li S, Kang X, Hu J. Image fusion with guided filtering[J]. IEEE T. Image Process., 2013, 22(7):2864-2875.
[13] Wang W, Chang F. A Multi-focus Image Fusion Method Based on Laplacian Pyramid[J]. J. Comput., 2011, 6(12):2559-2566.
[14] Kingsbury N. Image processing with complex wavelets[J]. Philos. Trans. R. Soc., A, 1999, 357(1760):2543-2560.
[15] Hariharan H, Koschan A, Abidi M. The direct use of curvelets in multifocus fusion[J] Proc. IEEE Int. Conf. Image Process., 2009, 2185-2188.
[16] Burt P J, Adelson E H. The laplacian pyramid as a compact image code[J]. IEEE T. Commun., 1983, 31(4):532-540.
[17] Toet A. Image fusion by a ratio of low-pass pyramid[J]. Pattern Recogn. Lett., 1989, 9(4):245-253.
[18] Chipman L J, Orr T M, Graham L N. Wavelets and image fusion[J]. Proc. IEEE Int. Conf. Image Process., 1995, 3:248-251.
[19] Lewis J J, O'Callaghan R J, Nikolov S G, Bull D R, Canagarajah N. Pixel-and region-based image fusion with complex wavelets[J]. Inform. Fusion, 2007, 8(2):119-130.
[20] Nencini F, Garzelli A, Baronti S, Alparone L. Remote sensing image fusion using the curvelet transform[J]. Inform. Fusion, 2007, 8(2):143-156.
[21] Naidu V. Image fusion technique using multi-resolution singular value decomposition[J]. Defence Sci. J., 2011, 61(5):479-484.
[22] Bredies K, Kunisch K, Pock T. Total generalized variation[J]. SIAM J. Imaging Sci., 2010, 3(3):492-526.
[23] Lysaker M, Lundervold A, Tai X C. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time[J]. IEEE T. Image Process., 2003, 12:1579-1590.
[24] Tai X C, Hahn J, Chung G J. A fast algorithm for Euler's Elastica model using augmented Lagrangian method[J]. SIAM J. Imaging Sci., 2011, 4(1):313-344.
[25] Zhu W, Tai X C. Image denoising using mean curvature of image surface[J]. SIAM J. Imaging Sci., 2012, 5(1):1-32.
[26] Chan T F, Golub G H, Mulet P. A nonlinear primal-dual method for total variation-based image restoration[J]. SIAM J. Sci. Comput., 1999, 20:1964-1977.
[27] Chambolle A. An algorithm for total variation minimization and applications[J]. J. Math. Imaging Vis., 2004, 20:89-97.
[28] Osher S, Burger M, Goldfarb D, Xu J, Yin W. An iterative regularization method for total variation-based image restoration[J]. Multiscale Model. Sim., 2005, 4:460-489.
[29] Wu C, Tai X C. Augmented Lagrangian method, dual methods, and split Bregman iteration for ROF, vectorial TV, and high order models[J]. SIAM J. Imaging Sci., 2010, 3:300-339.
[30] Han H, Huang Z. The Tailored finite point method[J]. Comput. Meth. Appl. Mat., 2014, 14:321-345.
[31] Marquina A, Osher S J. Image super-resolution by TV-regularization and Bregman iteration[J]. J. Sci. Comput., 2008, 37(3):367-382.
[32] Han H, Huang Z, Kellogg R B. A Tailored Finite Point Method for A Singular Perturbation Problem on An Unbounded Domain[J]. J. Sci. Comput., 2008, 36:243-261.
[33] Han H, Huang Z. Tailored finite point method based on exponential bases for convectiondiffusion-reaction equation[J]. Math. Comput., 2013, 82:213-226.
[34] Huang Z, Yang Y. Tailored finite point method for parabolic problems[J]. Comput. Meth. Appl. Mat., 2016, 16:543-562.
[35] Yang W, Huang Z, Zhu W. An Efficient Tailored Finite Point Method for Rician Denoising and Deblurring[J]. Commun. Comput. Phys., 2018, 24(4):1169-1195.
[36] Yang W, Huang Z, Zhu W. Image Segmentation Using the Cahn-Hilliard Equation[J]. J. Sci. Comput., 2019, 79(2):1057-1077.
[37] Huber P J. Robust estimation of a location parameter[J]. Ann. Math. Stat., 1964, 35(1):73-101.
[38] Nikolova M, Ng M K. Analysis of half-quadratic minimization methods for signal and image recovery[J]. SIAM J. Sci. Comput., 2005, 27(3):937-966.
[1] 马玉敏, 蔡邢菊. 求解带线性约束的凸优化的一类自适应不定线性化增广拉格朗日方法[J]. 计算数学, 2022, 44(2): 272-288.
阅读次数
全文


摘要