基于双树复小波变换与引导滤波的红外与可见光图像融合算法
|
摘要
针对传统的红外与可见光图像融合中多尺度融合方法存在图像细节丢失、边缘模糊以及对比度下降等不足,提出了基于双树复小波变换(DTCWT)和引导滤波器的红外与可见光图像融合方法。首先,利用DTCWT分解源图像,获取源图像的低频子带系数和高频子带系数;其次,对分解得到的低频系数基于区域能量加权融合,对高频系数先基于区域能量取大融合,再应用引导滤波进行细节增强;最后,通过DTCWT逆变换得到融合图像。实验结果表明,融合后的图像能够较好地保留细节信息,同时使边缘更加清晰,具有更好的视觉效果,且信息熵、互信息量、边缘相似度等客观评价指标相对传统方法均有不同程度的提高。
The traditional multi-scale fusion method of infrared-and-visible-image fusion has the disadvantages of image-detail loss, edge blur and contrast reduction. Therefore, an infrared-and-visibleimage fusion method based on DTCWT(dual-tree complex wavelet transform) and guided filter is proposed.Firstly, the DTCWT is used to decompose the source image and obtain the low frequency subband coefficient and the high frequency subband coefficient of the source image. Then based on the regional energy, the low-frequency coefficients obtained by the decomposition are weighted and fused, and firstly based on the regional energy, the high-frequency coefficients are fused, and then the guided filtering is applied to enhancing the detail. Finally, the fused image is obtained by inverse transform of DTCWT. The experimental results show that the fused image can retain the details better, and make the edges clearer and have better visual effects. In addition, the objective evaluation indexes such as information entropy, mutual information,and edge similarity are improved in varying degrees as compared with the traditional methods.
The traditional multi-scale fusion method of infrared-and-visible-image fusion has the disadvantages of image-detail loss, edge blur and contrast reduction. Therefore, an infrared-and-visibleimage fusion method based on DTCWT(dual-tree complex wavelet transform) and guided filter is proposed.Firstly, the DTCWT is used to decompose the source image and obtain the low frequency subband coefficient and the high frequency subband coefficient of the source image. Then based on the regional energy, the low-frequency coefficients obtained by the decomposition are weighted and fused, and firstly based on the regional energy, the high-frequency coefficients are fused, and then the guided filtering is applied to enhancing the detail. Finally, the fused image is obtained by inverse transform of DTCWT. The experimental results show that the fused image can retain the details better, and make the edges clearer and have better visual effects. In addition, the objective evaluation indexes such as information entropy, mutual information,and edge similarity are improved in varying degrees as compared with the traditional methods.
引文
[1]周渝人.红外与可见光图像融合算法研究[D].北京:中国科学院,2014.ZHOU Yu-ren.Infrared and Visible Image Fusion Algorithm[D].Beijing:Chinese Academy of Sciences,2014.
[2]Candes E J.Ridgelets:Theory and Application[D].Stanford:Stanford University,1998.
[3]Nenci F,Garzelli A,Baronti S.Remote Sensing Image Fusion Using the Curvelet Transform[J].Information Fusion,2007,8(02):143-156.
[4]Guo B L,Zhang Q,Hou Y.Region-based Fusion of Infrared and Visible Images Useing Nonsubsampled Contourlet Transform[J].Chinese Optics Letters,2008,6(05):338-341.
[5]Chen X G,Li X K.Application of Uniform Design and Genetic Algorithm in Optimization of Reversed Phase Chromatographic Separation[J].Chemometrics and Intelligent Laboratory Systems,2003,68(02):157-166.
[6]Li S T,Kang X D,Hu J W.Image Fusion with Guided Filtering[J].IEEE Transactions on Image Processi ng,2013,22(07):2864-2875.
[7]范艳玲,杨明.基于双树复小波变换的多聚焦图像融合[J].核电子学与探测技术,2014,34(02):204-208.FANG Yan-ling,YANG Ming.Multi-focus Image Fusion Based on Bi-tree Complex Wavelet Transform[J].Nuclear Electronics and Detection Technology,2014,34(02):204-208.
[8]Rama H,Velagapudi H.Guided Filter for Image Enhancement by Using the Contourlet Transform[J].International Journal of Scientific Engineering and Technology Research,2015,4(38):8348-8353.
[9]孙晓龙,王正勇,符耀庆等.基于改进拉普拉斯能量和的快速图像融合[J].计算机工程与应用,2015,51(05):193-197.SUN Xiao-long,WANG Zheng-yong,FU Yao-qing.Fast Image Fusion Based on Improved Laplace Energy Sum[J].Computer Engineering and Applicati on,2015,51(05):193-197.
[10]屈小波,闫敬文,杨贵德.改进拉普拉斯能量和的尖锐频率局部化Contourlet域多聚焦图像融合方法[J].光学精密工程,2009,17(05):1203-1212.QU Xiao-bo,YAN Jing-wen,YANG Gui-de.Improved Laplacian Energy Sum Localization of Sharp Frequency Multi-focus Image Fusion in Contourlet Domain[J].Optical Precision Engineering,2009,17(05):1203-1212.
[11]陈浩,王延杰.基于拉普拉斯金字塔变换的图像融合算法研究[J].激光与红外,2009,39(04):439-442.CHEN Hao,WANG Yan-jie.Image Fusion Algorithm Based on Laplacian Pyramid Transform[J].Laser and Infrared,2009,39(04):439-442.
[12]陈铭生,赖炬铭,孙季丰.基于离散多小波变换的医学图像融合[J].江西科学,2006,24(04):209-212.CHEN Ming-sheng,LAI Ju-ming,SUN Ji-feng.Medical Image Fusion Based on Discrete Multiwavelet Transform[J].Jiangxi Science,2006,24(04):209-212.
[13]隆刚,肖磊,陈学佺.Curvelet变换在图像处理中的应用综述[J].计算机研究与发展,2005,42(08):1331-1337.LONG Gang,XIAO Lei,CHEN Xue-quan.Application of Curvelet Transform in Image Processing[J].Computer Research and Development,2005,42(08):1331-1337.
[14]Robertsj W,Aardtj AV,Ahmed F B.Assessment of Image Fusion Procedures Using Entropy,Image Quality,and Multi Spectral Classification[J].Journal of Applied Remote Sensing,2008,2(01):023522.
[15]Jin H Y,Wang Y Y.A Fusion Method for Visible and Infrared Images Based on Contrast Pyramid with Teaching Learning Based Optimization[J].Infrared Physics&.Technology,2014(64):134-142.
[16]Eskicioglu A M,Fisher PS.Image Quality Measures and Their Performance[J].IEEE Transactions on Communicat ions,1996,43(12):2959-2965.
[2]Candes E J.Ridgelets:Theory and Application[D].Stanford:Stanford University,1998.
[3]Nenci F,Garzelli A,Baronti S.Remote Sensing Image Fusion Using the Curvelet Transform[J].Information Fusion,2007,8(02):143-156.
[4]Guo B L,Zhang Q,Hou Y.Region-based Fusion of Infrared and Visible Images Useing Nonsubsampled Contourlet Transform[J].Chinese Optics Letters,2008,6(05):338-341.
[5]Chen X G,Li X K.Application of Uniform Design and Genetic Algorithm in Optimization of Reversed Phase Chromatographic Separation[J].Chemometrics and Intelligent Laboratory Systems,2003,68(02):157-166.
[6]Li S T,Kang X D,Hu J W.Image Fusion with Guided Filtering[J].IEEE Transactions on Image Processi ng,2013,22(07):2864-2875.
[7]范艳玲,杨明.基于双树复小波变换的多聚焦图像融合[J].核电子学与探测技术,2014,34(02):204-208.FANG Yan-ling,YANG Ming.Multi-focus Image Fusion Based on Bi-tree Complex Wavelet Transform[J].Nuclear Electronics and Detection Technology,2014,34(02):204-208.
[8]Rama H,Velagapudi H.Guided Filter for Image Enhancement by Using the Contourlet Transform[J].International Journal of Scientific Engineering and Technology Research,2015,4(38):8348-8353.
[9]孙晓龙,王正勇,符耀庆等.基于改进拉普拉斯能量和的快速图像融合[J].计算机工程与应用,2015,51(05):193-197.SUN Xiao-long,WANG Zheng-yong,FU Yao-qing.Fast Image Fusion Based on Improved Laplace Energy Sum[J].Computer Engineering and Applicati on,2015,51(05):193-197.
[10]屈小波,闫敬文,杨贵德.改进拉普拉斯能量和的尖锐频率局部化Contourlet域多聚焦图像融合方法[J].光学精密工程,2009,17(05):1203-1212.QU Xiao-bo,YAN Jing-wen,YANG Gui-de.Improved Laplacian Energy Sum Localization of Sharp Frequency Multi-focus Image Fusion in Contourlet Domain[J].Optical Precision Engineering,2009,17(05):1203-1212.
[11]陈浩,王延杰.基于拉普拉斯金字塔变换的图像融合算法研究[J].激光与红外,2009,39(04):439-442.CHEN Hao,WANG Yan-jie.Image Fusion Algorithm Based on Laplacian Pyramid Transform[J].Laser and Infrared,2009,39(04):439-442.
[12]陈铭生,赖炬铭,孙季丰.基于离散多小波变换的医学图像融合[J].江西科学,2006,24(04):209-212.CHEN Ming-sheng,LAI Ju-ming,SUN Ji-feng.Medical Image Fusion Based on Discrete Multiwavelet Transform[J].Jiangxi Science,2006,24(04):209-212.
[13]隆刚,肖磊,陈学佺.Curvelet变换在图像处理中的应用综述[J].计算机研究与发展,2005,42(08):1331-1337.LONG Gang,XIAO Lei,CHEN Xue-quan.Application of Curvelet Transform in Image Processing[J].Computer Research and Development,2005,42(08):1331-1337.
[14]Robertsj W,Aardtj AV,Ahmed F B.Assessment of Image Fusion Procedures Using Entropy,Image Quality,and Multi Spectral Classification[J].Journal of Applied Remote Sensing,2008,2(01):023522.
[15]Jin H Y,Wang Y Y.A Fusion Method for Visible and Infrared Images Based on Contrast Pyramid with Teaching Learning Based Optimization[J].Infrared Physics&.Technology,2014(64):134-142.
[16]Eskicioglu A M,Fisher PS.Image Quality Measures and Their Performance[J].IEEE Transactions on Communicat ions,1996,43(12):2959-2965.