A Divide-and-Conquer Paradigm for Hyperspectral Classification and Target Recognition

Chapter
Part of the Augmented Vision and Reality book series (Augment Vis Real, volume 3)

Abstract

In this chapter, a multi-classifier, decision fusion framework is proposed for robust classification of high dimensional hyperspectral data in small-sample-size conditions. Such datasets present two key challenges. (1) The high dimensional feature spaces compromise the classifiers’ generalization ability in that the classifier tends to over-fit decision boundaries to the training data. This phenomenon is commonly known as the Hughes phenomenon in the pattern classification community. (2) The small-sample-size of the training data results in ill-conditioned estimates of its statistics. Most classifiers rely on accurate estimation of these statistics for modeling training data and labeling test data, and hence ill-conditioned statistical estimates result in poorer classification performance. Conventional approaches, such as Stepwise Linear Discriminant Analysis (S-LDA) are sub-optimal, in that they utilize a small subset of the rich spectral information provided by hyperspectral data for classification. In contrast, the approach proposed in this chapter utilizes the entire high dimensional feature space for classification by identifying a suitable partition of this space, employing a bank-of-classifiers to perform “local” classification over this partition, and then merging these local decisions using an appropriate decision fusion mechanism. Adaptive classifier weight assignment and nonlinear pre-processing (in kernel induced spaces) are also proposed within this framework to improve its robustness over a wide range of fidelity conditions. This chapter demonstrates the efficacy of the proposed algorithms to classify remotely sensed hyperspectral data, since these applications naturally result in very high dimensional feature spaces and often do not have sufficiently large training datasets to support the dimensionality of the feature space. Experimental results demonstrate that the proposed framework results in significant improvements in classification accuracies over conventional approaches.

Keywords

Decision Fusion Hyperspectral Imagery Kernel Discriminant Analysis Multi-Classifiers Small-Sample-Size Conditions Statistical Pattern Recognition 

References

  1. 1.
    Prasad, S., Bruce, L.M.: Decision fusion with confidence-based weight assignment for hyperspectral target recognition. IEEE Trans. Geosci. Remote Sens. 46(5), 1448–1456 (2008)CrossRefGoogle Scholar
  2. 2.
    Prasad, S., Bruce, L.M.: Information fusion in kernel-induced spaces for robust subpixel hyperspectral ATR. IEEE Geosci. Remote Sens. Lett. 6, 572–576 (2009)CrossRefGoogle Scholar
  3. 3.
    Duda, R.O., Hart, P.E., Stork, D.G.: Pattern Recognition, 2nd edn. Wiley-Interscience, Hoboken (2000)Google Scholar
  4. 4.
    Jain, A.K., Duin, R.P.W., Mao, J.: Statistical pattern recognition: a review. IEEE Trans. Pattern Anal. Mach. Intell. 22, 33 (2000)CrossRefGoogle Scholar
  5. 5.
    Farrell Jr, M.D., Mersereau, R.M.: On the impact of PCA dimension reduction for hyperspectral detection of difficult targets. IEEE Geosci. Remote Sens. Lett. 2, 192–195 (2005)CrossRefGoogle Scholar
  6. 6.
    Swets, D.L., Weng, J.J.: Using discriminant eigenfeatures for image retrieval. IEEE Trans. Pattern Anal. Mach. Intell. 18, 831–836 (1996)CrossRefGoogle Scholar
  7. 7.
    Prasad, S., Bruce, L.M.: Limitations of principal components analysis for hyperspectral target recognition. IEEE Geosci. Remote Sens. Lett. 5, 625–629 (2008)CrossRefGoogle Scholar
  8. 8.
    Prasad, S., Bruce, L.M.: Limitations of subspace LDA in hyperspectral target recognition applications. In: Proceedings of the IEEE Geoscience and Remote Sensing Symposium, pp. 4049–4052 (2007)Google Scholar
  9. 9.
    Prasad, S., Bruce, L.M.: Overcoming the small sample size problem in hyperspectral classification and detection tasks. In: Proceedings of the IEEE International Geoscience and Remote Sensing Symposium, pp. V-381–V-384 (2008)Google Scholar
  10. 10.
    Pu, R., Gong, P.: Band selection from hyperspectral data for conifer species identification. Presented at the Proceedings of the Geoinformatics’00 Conference, pp 139–146, June 2000Google Scholar
  11. 11.
    Cheriyadat, A., Bruce, L.M.: Why principal component analysis is not an appropriate feature extraction method for hyperspectral data. In: Proceedings of the IEEE Geoscience and Remote Sensing Symposium, vol. 6, pp. 3420–3422 (2003)Google Scholar
  12. 12.
    Kumar, S., Ghosh, J., Crawford, M.M.: Best-bases feature extraction algorithms for classification of hyperspectral data. IEEE Trans. Geosci. Remote Sens. 39, 1368–1379 (2001)CrossRefGoogle Scholar
  13. 13.
    Prasad, S., Bruce, L.M.: Hyperspectral feature space partitioning via mutual information for data fusion. In: Proceedings of the IEEE International Geoscience and Remote Sensing Symposium, pp. 4846–4849 (2007)Google Scholar
  14. 14.
    Tsagaris, V., Anastassopoulos, V., Lampropoulos, G.A.: Fusion of hyperspectral data using segmented PCT for color representation and classification. IEEE Trans. Geosci. Remote Sens. 43, 2365–2375 (2005)CrossRefGoogle Scholar
  15. 15.
    Cover, T.: Elements of Information Theory, 2nd edn. Wiley, New York (2006)MATHGoogle Scholar
  16. 16.
    Fukunaga, K.: Introduction to Statistical Pattern Recognition. Academic, New York (1990)MATHGoogle Scholar
  17. 17.
    Schlkopf, B., Smola, A.J.: Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond. The MIT Press, Cambridge (2001)Google Scholar
  18. 18.
    Mika, S., Ratsch, G., Weston, J., Scholkopf, B., Muller, K.-R.: Fisher discriminant analysis with kernels. In: Proceedings of IEEE Neural Networks for Signal Processing Workshop (1999)Google Scholar
  19. 19.
    Baudat, G., Anouar, F.: Generalized discriminant analysis using a kernel approach. In: Proceedings of Neural Computation (2000)Google Scholar
  20. 20.
    Benediktsson, J.A., Sveinsson, J.R.: Multisource remote sensing data classification based on consensus and pruning. IEEE Trans. Geosci. Remote Sens. 41, 932–936 (2003)CrossRefGoogle Scholar
  21. 21.
    Benediktsson, J.A., Swain, P.H.: Consensus theoretic classification methods. IEEE Trans. Syst. Man Cybern. 22, 688–704 (1992)CrossRefMATHGoogle Scholar
  22. 22.
    Fauvel, M., Chanussot, J., Benediktsson, J.A.: Decision fusion for the classification of urban remote sensing images. IEEE Trans. Geosci. Remote Sens. 44, 2828–2838 (2006)CrossRefGoogle Scholar
  23. 23.
    Analytical Spectral Devices FieldspecPro FR Specifications. http://asdi.com/productsspecifications-FSP.asp
  24. 24.
    SpecTIR ProSpecTIR-VINIR Sensor Specifications. http://www.spectir.com/
  25. 25.
    Ellis, J.M., Griffin, J.L., Vidrine, P.R., Godley, J.L.: Corn response to simulated drift of roundup ultra and liberty and utility of drift agents. Proc. South. Weed Sci. Soc. 51, 21 (1998)Google Scholar
  26. 26.
    Rowland, C.D.: Crop tolerance to non-target and labeled herbicide applications. M.S. Thesis, Mississippi State University, Mississippi State, MS (2000)Google Scholar
  27. 27.
    Ball, J.E.: Three stage level set segmentation of mass core, periphery, and spiculations for automated image analysis of digital mammograms. Ph.D. Dissertation, Department of Electrical Engineering, Mississippi State University, May 2007Google Scholar
  28. 28.
    Ball, J.E., West, T., Prasad, S., Bruce, L.M.: Level set hyperspectral image segmentation using spectral information divergence-based best band selection. In: Proceedings of the IEEE Geoscience and Remote Sensing Symposium, pp. 4053–4056 (2007)Google Scholar
  29. 29.
    HYPERION instrument specifications. http://eo1.gsfc.nasa.gov/Technology/Hyperion.html
  30. 30.
    Prasad, S.: Multi-classifiers and decision fusion for robust statistical pattern recognition with applications to hyperspectral classification. Ph.D. Dissertation, Department of Electrical and Computer Engineering, Mississippi State University (2008)Google Scholar
  31. 31.
    Kalluri, H., Prasad, S., Bruce, L.M.: Data dependant adaptation for improved classification of hyperspectral imagery. In: Proceedings of the IEEE Geoscience and Remote Sensing Symposium (IGARSS), Hawaii, USA (2010)Google Scholar
  32. 32.
    Kalluri, H., Prasad, S., Bruce, L.M.: Decision level fusion of spectral reflectance and derivative information for hyperspectral classification and target recognition. IEEE Trans. Geosci. Remote Sens. 48(11) 4047–4058 (2010)Google Scholar
  33. 33.
    Kalluri, H., Prasad, S., Bruce, L.M.: Fusion of spectral reflectance and derivative information for robust hyperspectral land cover classification. In: Proceedings of the IEEE Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, Grenoble, France (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  1. 1.Geosystems Research Institute and Electrical and Computer Engineering DepartmentMississippi State UniversityMississippi StateUSA

Personalised recommendations