Endmember Learning with KMeans through SCD Model in Hyperspectral Scene Reconstructions
Abstract
:1. Introduction
2. Prior Work in Dictionary Learning (DL)
3. Proposed Algorithm for Dictionary Learning (DL)
Algorithm 1 Proposed KMeans SCD (KMSCD) algorithm. 

Algorithm 2 Proposed for scene simulators: KMSCD+FNNOMP. 

4. Data Sets and Accuracy of Scene Reconstruction Assessments
5. Results
5.1. Feasibility of KMeans Clustering for Multispectral Data Set
5.2. CSCD vs. KMSCD: Reconstruction of Background Pixels
5.2.1. Robustness of CSCD and the Proposed KMSCD
5.2.2. Accuracy of CSCD and KMSCD: Background Pixels
5.3. Reconstruction of Trace Materials in the Scene
5.3.1. CSCD vs. KMSCD
5.3.2. KMSCD for Scene Simulation Applications
6. Discussion and Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
Abbreviations
ACE  Adaptive Cosine Estimator 
AUC  Area Under Curve 
CSCD  Classic Sparse Coding Dictionary 
DL  Dictionary Learning 
ED  Euclidean Distance 
EM  Endmember 
GSD  Ground Sampling Distance 
KMSCD  KMeans Sparse Coding Dictionary 
LUT  Lookup Table 
ROC  Receiver Operating Characteristic 
SCD  Sparse Coding Dictionary 
HSI  Hyperspectral Image 
MD  Manhattan Distance 
MSI  Multispectral Image 
ROC  Receiver Operating Characteristics 
TMM  Texture Material Mapper 
References
 Yuen, P.W.; Richardson, M. An introduction to hyperspectral imaging and its application for security, surveillance and target acquisition. Imaging Sci. J. 2010, 58, 241–253. [Google Scholar] [CrossRef]
 Ward, J.T. Realistic Texture in Simulated Thermal Infrared Imagery. Ph.D. Thesis, Chester F. Carlson Center for Imaging Science, Rochester Institute of Technology, Rochester, NY, USA, 2008. [Google Scholar]
 Pereira, W.; Richtsmeier, S.; Carr, S.; Kharabash, S.; Brady, A. A comparison of MCScene and CameoSim simulations of a real scene. In Proceedings of the 2014 6th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), Lausanne, Switzerland, 24–27 June 2014; pp. 1–4. [Google Scholar] [CrossRef]
 Evans, R. Modeling of SOC700 Hyperspectral Imagery with the CAMEOSIM Code. Proceedings of the 2007 Ground Systems Modeling, Validation & Testing Conference. 2007. Available online: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1000.5149&rep=rep1&type=pdf (accessed on 13 November 2019).
 James, I.; Richardson, M.; O’Keefe, E. Comparison of empirical and predicted ultraviolet aircraft signatures. Opt. Eng. 2019, 58, 025103. [Google Scholar] [CrossRef]
 Li, H.; Shen, H.; Yuan, Q.; Zhang, H.; Zhang, L.; Zhang, L. Quality improvement of hyperspectral remote sensing images: A technical overview. In Proceedings of the 2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), Los Angeles, CA, USA, 21–24 August 2016; pp. 1–5. [Google Scholar] [CrossRef]
 BioucasDias, J.M.; Plaza, A.; Dobigeon, N.; Parente, M.; Du, Q.; Gader, P.; Chanussot, J. Hyperspectral Unmixing Overview: Geometrical, Statistical, and Sparse RegressionBased Approaches. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2012, 5, 354–379. [Google Scholar] [CrossRef]
 Charles, A.S.; Olshausen, B.A.; Rozell, C.J. Learning Sparse Codes for Hyperspectral Imagery. IEEE J. Sel. Top. Signal Process. 2011, 5, 963–978. [Google Scholar] [CrossRef]
 Han, X.; Yu, J.; Luo, J.; Sun, W. Reconstruction From Multispectral to Hyperspectral Image Using Spectral LibraryBased Dictionary Learning. IEEE Trans. Geosci. Remote Sens. 2019, 57, 1325–1335. [Google Scholar] [CrossRef]
 Aharon, M.; Elad, M.; Bruckstein, A. KSVD: An algorithm for designing overcomplete dictionaries for sparse representation. IEEE Trans. Signal Process. 2006, 54, 4311–4322. [Google Scholar] [CrossRef]
 Fu, X.; Ma, W.; Chan, T.; BioucasDias, J.M. SelfDictionary Sparse Regression for Hyperspectral Unmixing: Greedy Pursuit and Pure Pixel Search Are Related. IEEE J. Sel. Top. Signal Process. 2015, 9, 1128–1141. [Google Scholar] [CrossRef]
 Degerickx, J.; Okujeni, A.; Iordache, M.D.; Hermy, M.; Van der Linden, S.; Somers, B. A Novel Spectral Library Pruning Technique for Spectral Unmixing of Urban Land Cover. Remote Sens. 2017, 9, 565. [Google Scholar] [CrossRef]
 Fu, X.; Ma, W.; BioucasDias, J.M.; Chan, T. Semiblind Hyperspectral Unmixing in the Presence of Spectral Library Mismatches. IEEE Trans. Geosci. Remote Sens. 2016, 54, 5171–5184. [Google Scholar] [CrossRef]
 Kapoor, A.; Singhal, A. A comparative study of KMeans, KMeans++ and Fuzzy CMeans clustering algorithms. In Proceedings of the 2017 3rd International Conference on Computational Intelligence Communication Technology (CICT), Ghaziabad, India, 9–10 Februay 2017; pp. 1–6. [Google Scholar] [CrossRef]
 Haut, J.M.; Paoletti, M.; Plaza, J.; Plaza, A. Cloud implementation of the Kmeans algorithm for hyperspectral image analysis. J. Supercomput. 2017, 73, 514–529. [Google Scholar] [CrossRef]
 Liu, Y.; Guo, Y.; Li, F.; Xin, L.; Huang, P. Sparse Dictionary Learning for Blind Hyperspectral Unmixing. IEEE Geosci. Remote Sens. Lett. 2019, 16, 578–582. [Google Scholar] [CrossRef]
 Nascimento, J.M.P.; Dias, J.M.B. Vertex component analysis: A fast algorithm to unmix hyperspectral data. IEEE Trans. Geosci. Remote Sens. 2005, 43, 898–910. [Google Scholar] [CrossRef]
 Li, J.; Agathos, A.; Zaharie, D.; BioucasDias, J.M.; Plaza, A.; Li, X. Minimum Volume Simplex Analysis: A Fast Algorithm for Linear Hyperspectral Unmixing. IEEE Trans. Geosci. Remote Sens. 2015, 53, 5067–5082. [Google Scholar] [CrossRef]
 Li, J.; BioucasDias, J.M.; Plaza, A.; Liu, L. Robust Collaborative Nonnegative Matrix Factorization for Hyperspectral Unmixing. IEEE Trans. Geosci. Remote Sens. 2016, 54, 6076–6090. [Google Scholar] [CrossRef]
 Gao, D.; Hu, Z.; Ye, R. SelfDictionary Regression for Hyperspectral Image SuperResolution. Remote Sens. 2018, 10, 1574. [Google Scholar] [CrossRef]
 Fang, L.; Li, S.; Kang, X.; Benediktsson, J.A. Spectral–Spatial Classification of Hyperspectral Images With a SuperpixelBased Discriminative Sparse Model. IEEE Trans. Geosci. Remote Sens. 2015, 53, 4186–4201. [Google Scholar] [CrossRef]
 Bao, C.; Ji, H.; Quan, Y.; Shen, Z. Dictionary Learning for Sparse Coding: Algorithms and Convergence Analysis. IEEE Trans. Pattern Anal. Mach. Intell. 2016, 38, 1356–1369. [Google Scholar] [CrossRef]
 Schnass, K. Average Performance of Orthogonal Matching Pursuit (OMP) for Sparse Approximation. IEEE Signal Process. Lett. 2018, 25, 1865–1869. [Google Scholar] [CrossRef]
 Yaghoobi, M.; Wu, D.; Davies, M.E. Fast NonNegative Orthogonal Matching Pursuit. IEEE Signal Process. Lett. 2015, 22, 1229–1233. [Google Scholar] [CrossRef]
 Chang, C. Adaptive Linear Spectral Mixture Analysis. IEEE Trans. Geosci. Remote Sens. 2017, 55, 1240–1253. [Google Scholar] [CrossRef]
 Jacobsson, J. Terrain Model Generation for the Infra Red Scene Simulation Software SensorVision(TM). Master’s Thesis, School of Computer Science and Engineering, Royal Institute of Technology, Stockholm, Sweden, 2005. [Google Scholar]
 Piper, J. A new dataset for analysis of hyperspectral target detection performance. In Proceedings of the HSI 2014, Hyperspectral Imaging and Applications Conference, Coventry UK, 15–16 October 2014. [Google Scholar]
 Bernstein, L.S.; Jin, X.; Gregor, B.; AdlerGolden, S.M. Quick atmospheric correction code: Algorithm description and recent upgrades. Opt. Eng. 2012, 51, 111719. [Google Scholar] [CrossRef]
 Nasrabadi, N.M. Hyperspectral Target Detection: An Overview of Current and Future Challenges. IEEE Signal Process. Mag. 2014, 31, 34–44. [Google Scholar] [CrossRef]
 Basener, W.F.; Allen, B.; Bretney, K. Geometry of statistical target detection. J. Appl. Remote Sens. 2017, 11, 015012. [Google Scholar] [CrossRef]
 Houlbrook, A.W.; Gilmore, M.A.; Moorhead, I.R.; Filbee, D.R.; Stroud, C.A.; Hutchings, G.; Kirk, A. Scene simulation for camouflage assessment. In Targets and Backgrounds VI: Characterization, Visualization, and the Detection Process; International Society for Optics and Photonics: Bellingham, WA, USA, 2000; Volume 4029. [Google Scholar] [CrossRef]
 Sun, K.; Li, Y.; Gao, J.; Wang, J.; Wang, J.; Xie, J.; Ding, N.; Sun, D. Simulation system of airborne FLIR searcher. In International Symposium on Optoelectronic Technology and Application 2014: Infrared Technology and Applications; Guina, M., Gong, H., Niu, Z., Lu, J., Eds.; International Society for Optics and Photonics: Bellingham, WA, USA, 2014; Volume 9300, pp. 480–485. [Google Scholar] [CrossRef]
 AI, Z.; Zhang, L.; Zhang, J. A practical method of texture segmentation and transformation for radar image simulation. In Proceedings of the 2012 IEEE International Conference on Computer Science and Automation Engineering (CSAE), Zhangjiajie, China, 25–27 May 2012; Volume 3, pp. 311–317. [Google Scholar] [CrossRef]
 Zhou, Q. Dynamic scene simulation technology used for infrared seeker. In International Symposium on Photoelectronic Detection and Imaging 2009: Advances in Infrared Imaging and Applications; Puschell, J., Gong, H., Cai, Y., Lu, J., Fei, J., Eds.; International Society for Optics and Photonics: Bellingham, WA, USA, 2009; Volume 7383, pp. 792–797. [Google Scholar] [CrossRef]
Hyperspectral Images  Lines  Samples  Bands  Spectral Range (in $\mathsf{\mu}$m) 

Selene H23 VNIR  3752  1600  160  0.41 to 1 
Selene H23 Dual  1876  380  448  0.41 to 2.5 
Paso RoblesMonterey  5115  741  224  0.36 to 2.5 
Virginia City 18071211  6349  320  178  0.4 to 2.45 
Virginia City 18071220  6758  320  178  0.4 to 2.45 
Virginia City 18071259  6904  320  178  0.4 to 2.45 
Hyperspectral Images  Proposed  CSCD Unmix  SDSOMP  CoNMF  MVSA  VCA 

Selene H23 VNIR  1.249  1.601  2.542  2.327  2.990  2.344 
Selene H23 Dual  0.282  0.405  1.376  0.606  0.986  1.319 
Paso RoblesMonterey  1.227  1.222  4.274  0.768  9.037  9.037 
Virginia City 18071220  0.054  0.110  0.858  0.155  2.574  2.572 
Virginia City 18071259  0.061  0.128  1.057  0.173  2.827  2.825 
Mean error  0.57  0.69  2.02  0.81  3.68  3.62 
± Std  ±0.61  ±0.68  ±1.42  ±0.89  ±3.1  ±3.08 
Enhanced reconstruction accuracy  
over 5 datasets w.r.t. KMSCD  20.64%  251.79%  40.24%  540.93%  529.9% 
Hyperspectral Images  Proposed  SCD Unmix  SDSOMP  CoNMF  MVSA  VCA 

Selene H23 VNIR  1.47  1.6  2.65  2.29  2.77  2.83 
Selene H23 Dual  2.33  2.51  3.94  2.69  3.63  5.46 
Paso RoblesMonterey  1.93  1.99  7.37  1.93  15.79  15.79 
Virginia City 18071220  0.25  0.3  0.94  0.25  0.84  0.99 
Virginia City 18071259  0.25  0.28  0.98  0.25  0.82  0.99 
Mean error  1.24  1.34  3.18  1.48  4.77  5.21 
± Std  ±0.96  ±1.01  ±2.66  ±1.16  ±6.28  ±6.19 
Enhanced reconstruction accuracy  
over 5 datasets w.r.t. KMSCD  7.22%  154.9%  18.94%  282.83%  318.4% 
Hyperspectral Images  Proposed  SCD Unmix  SDSOMP  CoNMF  MVSA  VCA 

Selene H23 VNIR  9.2e03  1.0e03  1.66e02  1.43e02  1.73e02  1.77e02 
Selene H23 Dual  5.2e03  5.6e03  8.8e02  6.0e03  8.1e03  1.22e02 
Paso RoblesMonterey  8.6e03  8.9e03  3.29e02  8.6e03  7.05e02  7.05e02 
Virginia City 18071220  1.4e03  1.7e03  5.3e03  1.4e03  4.7e03  5.6e03 
Virginia City 18071259  1.4e03  1.6e03  5.5e03  1.4e03  4.6e03  5.6e03 
Mean error  5.2e03  5.6e03  1.38e02  6.3e03  2.1e02  2.23e02 
± Std  ±3.8e03  ±3.9e03  ±1.16e02  ±5.4e03  ±2.81e02  ±2.74e02 
Enhanced reconstruction accuracy  
over 5 datasets w.r.t. KMSCD  7.75%  167.83%  22.87%  307.75%  332.56% 
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Chatterjee, A.; Yuen, P.W.T. Endmember Learning with KMeans through SCD Model in Hyperspectral Scene Reconstructions. J. Imaging 2019, 5, 85. https://doi.org/10.3390/jimaging5110085
Chatterjee A, Yuen PWT. Endmember Learning with KMeans through SCD Model in Hyperspectral Scene Reconstructions. Journal of Imaging. 2019; 5(11):85. https://doi.org/10.3390/jimaging5110085
Chicago/Turabian StyleChatterjee, Ayan, and Peter W. T. Yuen. 2019. "Endmember Learning with KMeans through SCD Model in Hyperspectral Scene Reconstructions" Journal of Imaging 5, no. 11: 85. https://doi.org/10.3390/jimaging5110085