Image Processing Systems Institute, Russian Academy of Sciences,
Samara State Aerospace University

PDF, 324 kB

Full text of article: Russian language.

DOI: 10.18287/0134-2452-2014-38-3-494-502

Pages: 494-502.

Abstract:
The present work is concerned with the problem of selecting the best hyperspectral image (HSI) classification algorithm. There are compared the following algorithms in our paper: decision tree using cross-validation function, decision tree C4.5 (C5.0), Bayesian classifier , maximum likelihood classifier, minimizing MSE classifier, including a special case - classification on conjugation, spectral angle mapper classifier(for mean vector and nearest neighbor) and support vector machine (SVM). There are presented experimental results of these algorithms for hyperspectral images received by AVIRIS satellite and during SpecTIR project.

Key words:
hyperspectral image, decision tree, C5.0, Bayes, MSE, conjugation classification, spectral angle mapper classification, SVM.

References:

1. Datt, B Preprocessing EO-1 Hyperion hyperspectral data to support the application of agricultural indexes / B. Datt, T.R. McVicar, T.G. Van Niel, [et al.] // IEEE Transactions On Geoscience And Remote Sensing. – 2003. – Vol. 41(6). – P. 1246-1259.
2. Metternicht, G.I. Remote sensing of soil salinity: potentials and constraints / G.I. Metternicht, J.A. Zinck // Remote Sensing of Environment. – 2003. – Vol. 85(1). – P. 1-20.
   
3. Landgrebe, D. Hyperspectral image data analysis / D. Landgrebe // IEEE Signal Processing Magazine. – 2002. – Vol. 19(1). – P. 17-28.
   
4. Gashnikov, M.V. Hierarchical grid interpolation for hyperspectral image compression / M.V. Gashnikov, N.I. Glumov // Computer Optics. – 2014. – Vol. 38, № 1. – P. 87-93. – (In Russian).
   
5. Li, J. Hyperspectral Image Segmentation Using a New Bayesian Approach With Active Learning / J. Li, M.J. Bioucas-Dias, A. Plaza // IEEE Transactions On Geoscience And Remote Sensing. – 2011. – Vol. 49(10). – P. 3947-3960.
   
6. Tarabalka, Yu. Spectral-Spatial Classification of Hyperspectral Imagery Based on Partitional Clustering Techniques / Yu. Tarabalka, J.A. Benediktsson, J. Chanussot // IEEE Transactions On Geoscience And Remote Sensing. – 2009. – Vol. 47(8). – P. 2973-2987.
   
7. Melgani, F. Classification of hyperspectral remote sensing images with support vector machines / F. Melgani, L. Bruzzone // IEEE Transactions On Geoscience And Remote Sensing. – 2004. – Vol. 42(8). – P. 1778-1790.
   
8. Camps-Valls, G. Kernel-based methods for hyperspectral image classification / G. Camps-Valls, L. Bruzzone // IEEE Transactions On Geoscience And Remote Sensing. – 2005. – Vol. 43(6). – P. 1351-1362.
   
9. Fauvel, M. Spectral and spatial classification of hyperspectral data using SVMs and morphological profiles / M. Fauvel, J.A. Benediktsson, J. Chanussot, J.R. Sveinsson // IEEE Transactions On Geoscience And Remote Sensing. – 2008. – Vol. 46(11). – P. 3804-3814.
   
10. Shafri, H. The Performance of Maximum Likelihood, Spectral Angle Mapper, Neural Network and Decision Tree Classifiers in Hyperspectral Image Analysis / H. Shafri, A. Suhaili, S. Mansor // Journal of Computer Science. – 2007. – Vol. 3(6). – P. 419-423.
    
11. Perumal, K. Supervised Classification Performance Of Multispectral Images / K. Perumal, R. Bhaskaran // Journal of Computing. – 2010. – Vol. 2(2). – P. 124-129.
    
12. Kopenkov, V.N. An algorithm for automatic construction of computational procedure of non-linear local image processing on the basis of hierarchical regression / V.N. Kopenkov, V.V. Myasnikov // Computer Optics. – 2012. – Vol. 36, № 2. – P. 257-265. – (In Russian).
    
13. Fodor I. A Survey Of Dimension Reduction Techniques / I. Fodor // Technical Report UCRL-ID-148494, Lawrence Livermore National Laboratory – 2002.
    
14. Introduction to Statistical Pattern Recognition / K. Fukunaga. – Academic Press, 1990. – 592 p.
    
15. Quinlan, J.R. Induction of Decision Trees /J.R. Quinlan // Machine Learning. – 1986. – Vol. 1(1). – P. 81-106.
    
16. Mathematical methods for machine learning аfrom examples (machine learning theory). Lecture course on machine learning [Electronic resource] / K.V. Vorontsov. – 2011. – 141 p. – (In Russian).
    
17. Computer Image Processing, Part II: Methods and algorithms / A.V. Chernov, V.M. Chernov, M.A. Chicheva, V.A. Fursov, M.V. Gash­nikov, N.I. Glumov, N.Yu. Ilyasova, A.G. Khra­mov, A.O. Korepanov, A.V. Kupriyanov, E.V. Myasnikov, V.V. Myasnikov, S.B. Popov, V.V. Sergeyev; ed. by V.A. Soifer. – VDM Verlag, 2009. – 584 p.
    
18. C4.5: Programs for Machine Learning / J.R. Quinlan. – San Mateo: Morgan Kaufmann Publishers Inc., 1993. – 302 p.
    
19. Wang, M. A Novel Hyperspectral Classification Method Based on C5.0 Decision Tree of Multiple Combined Classifiers / M. Wang, K. Gao, L. Wang, X. Miu // Proceedings of the 2012 4-th International Conference on Computational and Information Sciences (ICCIS 2012). – 2012. – P. 373-376.
    
20. Denisova, A.Yu. Anomaly detection for hyperspectral imaginary / A.Yu. Denisova, V.V. Myasnikov // Computer Optics. – 2014. – Vol. 38, № 2. –P. 287-296.
    
21. Fursov, V.А. Thematic classification of hyperspectral images using conjugacy indicator / V.A. Fursov, S.A. Bibikov, O.A. Baj­da // Computer Optics. – 2014. – Vol. 38, № 1. – P. 154-158
    
22. Sohn, Y. Supervised and unsupervised spectral angle classifiers / Y. Sohn, N.S. Rebello // Photogrammetric Engineering and Remote Sensing. – 2002. – Vol. 68(12). – P. 1271-1280.
    
23. Pattern recognition. Mathematical methods. Program system. Practical use / Ju.I. Juravlev, V.V. Ryazanov, O.V. Senko. – Moscow: “Fazis” Publisher, 2005. – 159 p. – (In Russian).
24. Cortes, C. Support-Vector Networks / C. Cortes, V. Vapnik // Machine Learning. – 1995. – Vol. 20(3). – P. 273-297.

---

© 2009, IPSI RAS
151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: journal@computeroptics.ru ; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20