基于WNMF和区域分维的图像融合算法
|
摘要
针对非负矩阵分解图像融合算法细节表现能力不足的缺陷,提出了一种基于加权非负矩阵分解和区域分维相结合的红外与可见光图像融合算法。在研究图像区域分维性质的基础上,用不同尺度上的区域分维来获取加权系数。通过设计加权系数的获取方法,重点突出边缘像素和小区域,以提高加权非负矩阵分解图像融合算法的细节提取能力,并得到最符合人眼视觉效果的融合图像。与现有基于标准或各种改进非负矩阵分解图像融合算法的对比实验表明,所提算法在平均梯度等表示细节信息的指标上提高了19%以上,有效改善了标准非负矩阵分解图像融合算法存在的不足。
The non-negative matrix factorization based fusion algorithm can not extract details from source images effectively.In order to improve this defect of non-negative matrix factorization,a novel image fusion algorithm for infrared and visible images based on weighted non-negative matrix factorization and regional fractal dimension is proposed.The properties of regional fractal dimensions are researched,and the weighted coefficients are obtained through regional fractal dimensions on different scales.The weighted coefficients are designed for emphasizing the edges and small areas of the source images,so the fusion result that is more comfortable to observe and contains more details is obtained.Compared with other methods based on traditional non-negative matrix factorization,the proposed algorithm improves the visual effect and the average gradient of the fusion result is improved by more than 19%.
The non-negative matrix factorization based fusion algorithm can not extract details from source images effectively.In order to improve this defect of non-negative matrix factorization,a novel image fusion algorithm for infrared and visible images based on weighted non-negative matrix factorization and regional fractal dimension is proposed.The properties of regional fractal dimensions are researched,and the weighted coefficients are obtained through regional fractal dimensions on different scales.The weighted coefficients are designed for emphasizing the edges and small areas of the source images,so the fusion result that is more comfortable to observe and contains more details is obtained.Compared with other methods based on traditional non-negative matrix factorization,the proposed algorithm improves the visual effect and the average gradient of the fusion result is improved by more than 19%.
引文
[1]苗启广,王宝树.图像融合的非负矩阵分解算法[J].计算机辅助设计与图形学学报,2005,17(9):2029-2032.MIAO Q G,WANG B SH.A novel algorithm of multi-sensor image fusion using non-negative matrix factoriza-tion[J].Journal of Computer-Aided Design&ComputerGraphics,2005,17(9):2029-2032.
[2]孙锐,高隽.组合NMF和PCA的图像哈希方法[J].电子测量与仪器学报,2009,23(5):52-57.SUN R,GAO J.Image Hashing method via combinationof NMF and PCA[J].Journal of Electronic Measurementand Instrument,2009,23(5):52-57.
[3]王刚,马美仲,赵英路,等.Curvelet变换域图像融合算法[J].仪器仪表学报,2008,29(9):1841-1845.WANG G,MA M ZH,ZHAO Y L,et al.Algorithm forimage fusion in the Curvelet transform domain[J].Chi-nese Journal of Scientific Instrument,2008,29(9):1841-1845.
[4]Lewis J J,O′Callaghan R J,Nikolov S G,et al.Pixel-and region-based image fusion with complex wavelets[J].Information Fusion,2007,8(2):119-130.
[5]蒋年德,王耀南,毛建旭.基于Curvelet变换的遥感图像融合研究[J].仪器仪表学报,2008,29(1):61-66.JIANG N D,WANG Y N,MAO J X.Research on re-mote sensing image fusion based on Curvelet transform[J].Chinese Journal of Scientific Instrument,2008,29(1):61-66.
[6]苗启广,王宝树.基于非负矩阵分解的多聚焦图像融合研究[J].光学学报,2005,25(6):755-759.MIAO Q G,WANG B SH.Multi-focus image fusionbased on non-negative matrix factorization[J].Acta Op-tica Sinica,2005,25(6):755-759.
[7]YE Y S,ZHAO B J,TANG L B.SAR and visible imagefusion based on local non-negative matrix factorization[C].The Ninth International Conference on ElectronicMeasurement&Instruments,Beijing,2009:263-266.
[8]HUANG Z,YU X C,WANG G A,et al.Application ofseveral non-negative matrix factorization-based methods inremote sensing image fusion[C].Fifth International Con-ference on Fuzzy Systems and Knowledge Discovery,Tianjin,2008:29-33.
[9]ZHANG S W,CHEN J,MIAO D D.An image fusionmethod based on WNMF and region segmentation[C].2008 IEEE Pacific-Asia Workshop on Computational In-telligence and Industrial Application,Wuhan,2008:282-285.
[10]陈颙,陈凌.分形几何学[M].北京:地震出版社,2005,59-61.CHEN Y,CHEN L.Fractal geometry[M].Beijing:Earthquake Publishing House,2005,59-61.
[11]YU TH,LI Q M,DAI J M.Newenhancement of infra-red image based on human visual system[J].ChineseOptics Letters,2009,7(3):206-209.
[12]DANIEL D L,H SEBASTIAN SEUNG.Algorithms fornon-negative matrix factorization[C].Advances in Neu-ral Information Processing Systems,Denver,2000:556-562.
[13]李乐,章毓晋.非负矩阵分解算法综述[J].电子学报,2008,36(4):737-743.LI L,ZHANG Y J.A survey on algorithms of non-nega-tive matrix factorization[J].Acta Electronica Sinica,2008,36(4):737-743.
[14]ALEX P P.Fractal-based description of natural scenes[J].IEEE Transactions on Pattern Analysis and Ma-chine Intelligence,1984,6(6):661-674.
[15]ZHOU W,BOVIK A C.A universal image quality index[J].IEEE Signal Processing Letters,2002,9(3):81-84.
[2]孙锐,高隽.组合NMF和PCA的图像哈希方法[J].电子测量与仪器学报,2009,23(5):52-57.SUN R,GAO J.Image Hashing method via combinationof NMF and PCA[J].Journal of Electronic Measurementand Instrument,2009,23(5):52-57.
[3]王刚,马美仲,赵英路,等.Curvelet变换域图像融合算法[J].仪器仪表学报,2008,29(9):1841-1845.WANG G,MA M ZH,ZHAO Y L,et al.Algorithm forimage fusion in the Curvelet transform domain[J].Chi-nese Journal of Scientific Instrument,2008,29(9):1841-1845.
[4]Lewis J J,O′Callaghan R J,Nikolov S G,et al.Pixel-and region-based image fusion with complex wavelets[J].Information Fusion,2007,8(2):119-130.
[5]蒋年德,王耀南,毛建旭.基于Curvelet变换的遥感图像融合研究[J].仪器仪表学报,2008,29(1):61-66.JIANG N D,WANG Y N,MAO J X.Research on re-mote sensing image fusion based on Curvelet transform[J].Chinese Journal of Scientific Instrument,2008,29(1):61-66.
[6]苗启广,王宝树.基于非负矩阵分解的多聚焦图像融合研究[J].光学学报,2005,25(6):755-759.MIAO Q G,WANG B SH.Multi-focus image fusionbased on non-negative matrix factorization[J].Acta Op-tica Sinica,2005,25(6):755-759.
[7]YE Y S,ZHAO B J,TANG L B.SAR and visible imagefusion based on local non-negative matrix factorization[C].The Ninth International Conference on ElectronicMeasurement&Instruments,Beijing,2009:263-266.
[8]HUANG Z,YU X C,WANG G A,et al.Application ofseveral non-negative matrix factorization-based methods inremote sensing image fusion[C].Fifth International Con-ference on Fuzzy Systems and Knowledge Discovery,Tianjin,2008:29-33.
[9]ZHANG S W,CHEN J,MIAO D D.An image fusionmethod based on WNMF and region segmentation[C].2008 IEEE Pacific-Asia Workshop on Computational In-telligence and Industrial Application,Wuhan,2008:282-285.
[10]陈颙,陈凌.分形几何学[M].北京:地震出版社,2005,59-61.CHEN Y,CHEN L.Fractal geometry[M].Beijing:Earthquake Publishing House,2005,59-61.
[11]YU TH,LI Q M,DAI J M.Newenhancement of infra-red image based on human visual system[J].ChineseOptics Letters,2009,7(3):206-209.
[12]DANIEL D L,H SEBASTIAN SEUNG.Algorithms fornon-negative matrix factorization[C].Advances in Neu-ral Information Processing Systems,Denver,2000:556-562.
[13]李乐,章毓晋.非负矩阵分解算法综述[J].电子学报,2008,36(4):737-743.LI L,ZHANG Y J.A survey on algorithms of non-nega-tive matrix factorization[J].Acta Electronica Sinica,2008,36(4):737-743.
[14]ALEX P P.Fractal-based description of natural scenes[J].IEEE Transactions on Pattern Analysis and Ma-chine Intelligence,1984,6(6):661-674.
[15]ZHOU W,BOVIK A C.A universal image quality index[J].IEEE Signal Processing Letters,2002,9(3):81-84.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心