基于双图正则的半监督NMF混合像元解混算法

doi:10.11896／j.issn.1002-137X.2018.12.041

摘要/Abstract

摘要： 在高光谱图像中混合像元普遍存在,这极大地阻碍了高光谱遥感技术的发展进程,因此,在利用光谱图像的过程中,如何准确高效地进行混合像元解混是一个关键问题。对于高光谱图像混合像元分解,使用原始的非负矩阵分解(Nonnegative Matrix Factorization,NMF)算法面临一些困难:首先,其目标函数为非凸函数,难以求解得到全局最优解;其次,混合像元中并不存在纯像元。为了解决这些问题,文中提出一种新的算法——基于双图正则的半监督NMF(Dual graph-regularized Constrained Nonnegative Matrix Factorization,DCNMF)混合像元解混算法。该算法采用了梯度下降法和迭代更新法则,既考虑了高光谱数据流形与光谱特征流形的几何结构,又能跳出局部极值,从而求解得到全局最优解。通过真实的高光谱图像数据仿真实验表明,DCNMF算法能够准确高效地进行混合像元分解,改善了解混效果,提高了解混精度,节约了计算时间,加快了收敛速度。

关键词: 非负矩阵分解, 高光谱图像, 混合像元解混, 双图正则

Abstract: In hyperspectral images,the existence of mixed pixels greatly impedes the development of hyperspectral remote sensing technology.Therefore,how to carry out unmixing accurately and efficiently in the process of using spectral images is a key problem.For hyperspectral unmixing,using original non-negative matrix factorization (NMF) algorithm faces some difficulties,for example,the objective function is non-convex function,so it is difficult to solve the global optimal solution.Besides,the pure pixel like element doesn’t exist in mixed pixel.In order to solve these problems,this paper proposed a mixed pixel unmixing algorithm namely dual graph-regularized constrained semi-supervised NMF (DCNMF) .This algorithm adopts gradient descent algorithm and iterative updating rule,considers the geometric structures of hyperspectral data manifold and the spectral feature manifold,and can jump out of the local extremum,thus solving the global optimal solution.Real hyperspectral image data simulation experiments show that DCNMF algorithm can be used to decompose the mixed pixel accurately and efficiently,enhancing the effect of unmixing,improving the accuracy of mixing,saving the computing time and speeding up convergence.

Key words: Bigraph regularization, Hyperspectral images, Mixed pixel disintegration, Nonnegative matrix factorization

中图分类号:

TP391

邹丽, 蔡希彪, 孙静, 孙福明. 基于双图正则的半监督NMF混合像元解混算法[J]. 计算机科学, 2018, 45(12): 251-254. https://doi.org/10.11896／j.issn.1002-137X.2018.12.041

ZOU Li, CAI Xi-biao, SUN Jing, SUN Fu-ming. Hyperspectral Unmixing Algorithm Based on Dual Graph-regularized Semi-supervised NMF[J]. Computer Science, 2018, 45(12): 251-254. https://doi.org/10.11896／j.issn.1002-137X.2018.12.041

参考文献

[1]GOETZ A F,VANE G,SOLOMON J E,et al.Imaging spec-trometry for earth remote sensing[J].Science,1985,228(4704):1147-1153.
[2]VANE G,GREEN R,CHRIEN T G,et al.The airborne visible/infrared imaging spectrometer(AVIRIS)[J].Remote Sensing of Environment,1993,44(2):127-143.
[3]GREEN R,EASTWOOD M L,SARTURE C M,et al.Imaging spectroscopy and the airborne visible/infrared imaging spectrometer(AVIRIS)[J].Remote Sensing of Environment,1998,65(3):227-248.
[4]LEE D D,SEUNG H S.Learning the parts of objects by non-ne-gative matrix factorization[J].Nature,1999,401(6755):788-791.
[5]TONG L,ZHOU J,QIAN Y T.Nonnegative Matrix Factorization Based Hyperspectral Unmixing with Partially Known Endmembers[J].IEEE Transactions on Geoscience and Remote Sensing,2016,54(11):6531-6544.
[6]YU Y,GUO S,SUN W D.Minimum distance constrained nonnegative matrix factorization for the endmember extraction of hyperspectral images[C]∥Proceeding of Remote Sensing and GIS Data Processing and Applications.Wuhan,2007:6790151-6790159.
[7]JIA S,QIAN Y T,JI X,et al.Hyperspectral Unmixing Algorithm Based on spectral and spatial characteristics[J].Journal of Shenzhen University(Science & Engineering),2009,26(3):162-167.(in Chinese)
贾森,钱沄涛,纪霞,等.基于光谱和空间特性的高光谱解混方法[J].深圳大学学报(理工版),2009,26(3):162-167.
[8]YANG S Y,ZHANG X T,YAO Y G,et al.Geometric Nonnegative Matrix Factorization (GNMF) for Hyperspectral Unmixing[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sening,2015,8(6):2696-2703.
[9]WANG W H,QIAN Y T,TANG Y Y.Hypergraph-Regularized Sparse NMF for Hyperspectral Unmixing[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sening,2016,9(2):681-694.
[10]YUAN Y,FU M,LU X Q.Substance Dependence Constrained Sparse NMF for Hyperspectral Unmixing[J].IEEE Transactions on Geoscience and Remote Sensing,2015,53(6):2975-2986.
[11]ADAMS J B,SABOL D E,KAPOS V,et al.Classification of multispectral images based on fractions of endmembers:Application to land-cover change in the Brazilian Amazon[J].Remote Sensing of Environment,1995,52(2):137-154.
[12]ADAMS J B,SMITH M O,JOHNSON P E.Spectral mixture modeling:of rock and soil types at the Viking Larder 1 site[J].Journal of Geophysical Research:Solid Earth(1978-2012),1986,91(8):8098-8112.
[13]DIAS J B,PLAZA A.Hyperspectral unmixing geometrical,statistical and sparse regression-based approaches[C]∥Procee-dings of SPIE:Image and Signal Processing for Remote Sensing XVI.Toulouse,France:SPIE Press,2010.
[14]ZHAO C H,CHENG B Z,YANG W C.A hyperspectral unmi-xing algorithm based on the constraint nonnegative matrix decomposition[J].Journal of Harbin Institute of Technology,2012,33(3):378-382.(in Chinese)
赵春晖,成宝芝,杨伟超.利用约束非负矩阵分解的高光谱解混算法.哈尔滨工业大学学报,2012,33(3):378-382.
[15]SONG Y G,WU Z B,WEI Z H,et al.Survey of sparsity constrained hyperspectral unmixing.Journal of Nanjing University of Science and Technoloogy,2013,37(4):486-492.(in Chinese)
宋义刚,吴泽彬,韦志辉,等.稀疏性高光谱解混方法研究[J].南京理工大学学报,2013,37(4):486-492.
[16]KONG F J,BIAN C D,LI Y S,et al.Hyperspectral unmixing method for non-convex and low rank constraints[J].Journal of Xi’an Electronic and Science University(Natural Science Edition),2016,43(6):116-121.(in Chinese)
孔繁锵,卞陈鼎,李云松,等.非凸稀疏低秩约束的高光谱解混方法.西安电子科技大学学报(自然科学版),2016,43(6):116-121.
[17]WANG T C,LIU X Z,DONG Z Z,et al.An adaptive robust minimum volume hyperspectral unmixing algorithm[J].Journal of Automation,2017,43(2):1-19.(in Chinese)
王天成,刘相振,董泽政,等.一种自适应鲁棒最小体积高光谱解混算法[J].自动化学报,2017,43(2):1-19.
[18]LIU H,WU Z,CAI D,et al.Constrained non-negative matrix factorization for image representation[J].IEEE Transactions Pattern Analysis and Machine Intelligence,2012,34(7):1299-1311.
[19]SHU Z Q,ZHAO C X.Constrained nonnegative matrix decomposition algorithm based on graph regularization and its application in image representation[J].Pattern Recognition and Artificial Intelligence,2013,26(3):300-306.(in Chinese)
舒振球,赵春霞.基于图正则化的受限非负矩阵分解算法及其在图像表示中的应用[J].模式识别与人工智能,2013,26(3):300-306.
[20]PLAZA A,MARTINEZ P,PEREZ R,et al.A quantitative and comparative analysis of endmember extraction algorithms from hyperspectral data[J].IEEE Geoscience and Remote Sensing Letters,2004,42(3):650-663.
[21]KESHAVA N,MUSTARD J F.Spectral unmixing[J].IEEESignal Process Mag,2002,19(1):44-57.
[22]LANDGREBE D.Multispectral data analysis:a signal theoryperspective[D].West Lafayette:Purdye University,1998.
[23]SWAYZE G.The hydrothermal l and structural history of the cuprite mining district,southwestern Nevada:an integrated geological and geophysical approach[D].Boulde:University of Co-lorado,1997.

相关文章 15

[1]	赵亮, 张洁, 陈志奎. 基于双图正则化的自适应多模态鲁棒特征学习 Adaptive Multimodal Robust Feature Learning Based on Dual Graph-regularization 计算机科学, 2022, 49(4): 124-133. https://doi.org/10.11896/jsjkx.210300078
[2]	官铮, 邓扬琳, 聂仁灿. 光谱重建约束非负矩阵分解的高光谱与全色图像融合 Non-negative Matrix Factorization Based on Spectral Reconstruction Constraint for Hyperspectral and Panchromatic Image Fusion 计算机科学, 2021, 48(9): 153-159. https://doi.org/10.11896/jsjkx.200900054
[3]	郑建炜, 黄娟娟, 秦梦洁, 徐宏辉, 刘志. 基于非局部相似及加权截断核范数的高光谱图像去噪 Hyperspectral Image Denoising Based on Non-local Similarity and Weighted-truncated NuclearNorm 计算机科学, 2021, 48(9): 160-167. https://doi.org/10.11896/jsjkx.200600135
[4]	陶星朋, 徐宏辉, 郑建炜, 陈婉君. 基于非凸低秩矩阵逼近和全变分正则化的高光谱图像去噪 Hyperspectral Image Denoising Based on Nonconvex Low Rank Matrix Approximation and TotalVariation Regularization 计算机科学, 2021, 48(8): 125-133. https://doi.org/10.11896/jsjkx.200400143
[5]	段菲, 王慧敏, 张超. 面向数据表示的Cauchy非负矩阵分解 Cauchy Non-negative Matrix Factorization for Data Representation 计算机科学, 2021, 48(6): 96-102. https://doi.org/10.11896/jsjkx.200700195
[6]	巫勇, 刘永坚, 唐瑭, 王洪林, 郑建成. 基于鲁棒低秩张量恢复的高光谱图像去噪 Hyperspectral Image Denoising Based on Robust Low Rank Tensor Restoration 计算机科学, 2021, 48(11A): 303-307. https://doi.org/10.11896/jsjkx.210200103
[7]	李雨蓉, 刘杰, 刘亚林, 龚春叶, 王勇. 面向语音分离的深层转导式非负矩阵分解并行算法 Parallel Algorithm of Deep Transductive Non-negative Matrix Factorization for Speech Separation 计算机科学, 2020, 47(8): 49-55. https://doi.org/10.11896/jsjkx.190900202
[8]	李向利, 贾梦雪. 基于预处理的超图非负矩阵分解算法 Nonnegative Matrix Factorization Algorithm with Hypergraph Based on Per-treatments 计算机科学, 2020, 47(7): 71-77. https://doi.org/10.11896/jsjkx.200200106
[9]	王燕, 王丽. 面向高光谱图像分类的局部Gabor卷积神经网络 Local Gabor Convolutional Neural Network for Hyperspectral Image Classification 计算机科学, 2020, 47(6): 151-156. https://doi.org/10.11896/jsjkx.190500147
[10]	王丽星, 曹付元. 基于Huber损失的非负矩阵分解算法 Huber Loss Based Nonnegative Matrix Factorization Algorithm 计算机科学, 2020, 47(11): 80-87. https://doi.org/10.11896/jsjkx.190900144
[11]	周昌, 李向利, 李俏霖, 朱丹丹, 陈世莲, 蒋丽榕. 基于余弦相似度的稀疏非负矩阵分解算法 Sparse Non-negative Matrix Factorization Algorithm Based on Cosine Similarity 计算机科学, 2020, 47(10): 108-113. https://doi.org/10.11896/jsjkx.190700112
[12]	张显,叶军. 基于非局部相似联合低秩表示的高光谱图像去噪 Hyperspectral Images Denoising Based on Non-local Similarity Joint Low-rank Representation 计算机科学, 2020, 47(1): 170-175. https://doi.org/10.11896/jsjkx.181202337
[13]	康林瑶, 唐兵, 夏艳敏, 张黎. 基于GPU加速和非负矩阵分解的并行协同过滤推荐算法 GPU-accelerated Non-negative Matrix Factorization-based Parallel Collaborative Filtering Recommendation Algorithm 计算机科学, 2019, 46(8): 106-110. https://doi.org/10.11896/j.issn.1002-137X.2019.08.017
[14]	何孝文, 胡一飞, 王海平, 陈默. 在线学习非负矩阵分解 Online Learning Nonnegative Matrix Factorization 计算机科学, 2019, 46(6A): 473-477.
[15]	黄梦婷, 张灵, 姜文超. 基于流形正则化的多类型关系数据联合聚类方法 Multi-type Relational Data Co-clustering Approach Based on Manifold Regularization 计算机科学, 2019, 46(6): 64-68. https://doi.org/10.11896/j.issn.1002-137X.2019.06.008

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed