A method for hyperspectral satelite image classification using dempster’s combination rule

1Popov, MO, 1Podorvan, VM, 2Alpert, SІ
1State institution «Scientific Centre for Aerospace Research of the Earth of the Institute of Geological Sciences of the National Academy of Sciences of Ukraine», Kyiv, Ukraine
2State institution «Scientific Centre for Aerospace Research of the Earth Institute of Geological Science National Academy of Sciences of Ukraine», Kyiv, Ukraine
Kosm. nauka tehnol. 2015, 21 ;(1):25–37
https://doi.org/10.15407/knit2015.01.025
Section: Remote sensing of the Earth
Publication Language: Ukrainian
Abstract: 

This paper proposes the new method for hyperspectral satellite image classification using the Dempster-Shafer evidence theory. The method differs from the known solutions by introducing a procedure of assessing classification value of the spectral bands by a special empirical function and by specific approach to partition of the spectral feature space. Such an approach allows to use only the most informative spectral bands that significantly reduces the dimension of the feature space while improving classification accuracy. The software for the proposed method and the results of evaluation of the classification accuracy, which is really achieved, are presented.

Keywords: Dempster-Shafer’s Evidence Theory, hyperspectral satellite image, image classification, informativity function.
References: 

1. Al'pert S. I.  Assessment of the quality classification of images based on the matrix and the accuracy coefficients.  Mathematical machines and systems, No. 1, 101—107 (2014) [in Ukrainian].

2. Lyalko V. I., Popov M. O. (Eds.) Multispectral remote sensing in nature management, 360 p. (Nauk.dumka, Kyiv, 2006) [in Ukrainian].

3. Burshtinska Kh.V., Stankevich S.A. Aerospace imaging system.  316 p. (Lviv Polytechnic, Lviv, 2013) [in Ukrainian].

4. Venttsel' E. S. Probability theory (Teorija verojatnostej). 576 p. (Nauka, M., 1969) [in Russian].

5. Dulicheva Yu.Yu.  About Filtering Problems of Training Sample.  Artificial Intelligence. No. 2, 65—71 (2006) [in Russian].

6. Kozoderov V. V., Kondranin T. V., Kazancev O. Ju. et al. Processing and interpretation of aerospace hyperspectral measurements for remote diagnostics of natural and man-made objects.  Earth Res. from Space, No. 2, 36—54 (2009) [in Russian].

7. Popov M. A. Methodology for evaluating the accuracy of the classification of objects in space images.  Journal of Automation and Information Sciences, No.1, 97—103 (2007) [in Russian].

8. Popov M. O., Stankevytch S. A., Moldovan V. D. The role of  hyperspectral aerospace information in detection and tracking of objects.  Science and Defense, No.3, 25—35 (2006) [in Ukrainian].

9. Popov M. O., Topol'nic'kij M. V.  Objects classification on multispectral / hyperspectral aerospace images based on the Dempster-Shafer theory of evidence. Mathematical machines and systems, No.1, 58—69 (2014) [in Russian].

10. Stankevich S.A  Algorithm of statistical classifications of object’s remote monitoring of their spectral topological characteristics. Scientific Bulletin of National Mining University, No. 7, 38—40 (2006) [in Ukrainian].

11. Beynon M. J., Curry B., Morgan P. The Dempster — Shafer theory of evidence: an alternative approach to multicriteria decision modeling.  Omega  28(1), 37—50 (2000).
https://doi.org/10.1016/S0305-0483(99)00033-X

12. Bongasser M., Hungate W. S., Watkins R. Hyperspectral remote sensing: Principles and applications, 119 p. (CRC Press, Boca Raton, 2008).

13. Chang C.-I. Hyperspectral data processing: Algorithm design and analysis. 1164 p. (John Willey and Sons, Hoboken, N. J., 2013.).

14. Congalton G., Green K. Assessing the accuracy of remotely sensed data: Principles and practices. 2nd Ed. 183 p. (CRC Press, Boca Raton, 2009).

15. Dash M., Liu H. Feature selection for classification. Intel. Data An.  1, 131—156 (1997).

16. Gong P. Integrated analysis of spatial data from multiple sources: Using evidential reasoning and artificial neural network techniques for geological mapping.  Photogramm. Eng. and Remote Sens. 62, (5), 513—523 ( 1996).

17. Chang Ch.-I. (Ed.) Hyperspectral data exploitation: Theory and applications. 430 p. (John Willey and Sons, Hoboken, N. J., 2007).

18. Van der Meer F. D., de Jong S. M. (Eds.) Imaging spectrometry: Basic principles and prospective applications. 404 p. (Dodrecht: Kluwer, 2001).

19. Lein J. K. Applying evidential reasoning methods to agricultural land cover classification. Int. J. Remote Sens. 24 (21), 4161—4180 (2003).

20. Mertikas P., Zervakis M. E. Exemplifying the theory of evidence in remote sensing image classification. Int. J. Remote Sens.  22 (6), 1081—1095 (2001).

21. Popov M. A., Topolnitskiy M. V. A Dempster — Shafer evidence theory-based approach to object classification on multispectral / hyperspectral images. Proceedings of the 10th Int. Conf. IEEE on Digital Technologies (DT’2014). P. 296—300 (Žilina, 2014).

22. Shafer G.  A Mathematical Theory of Evidence. 297 p. (Princeton Univ. Press, Princeton, 1976).

23. Taroun A., Yang J. B. Dempster-Shafer theory of evidence: Potential usage for decision making and risk analysis in construction project management. The Built & Hum. Environ. Rev. 4 (1), 155—166 (2011).

24. Tso B. Classification methods for remotely sensed data. 332 p. (Tailor and Francis, London, 2001).

25. Varshey P. K., Arora M. K. Advanced image processing techniques for remotely sensed hyperspectral data. 322 p. (Springer-Verlag, Berlin-Heidelberg, 2004).