Generating labeled samples for hyperspectral image classification using correlation of spectral bands

Abstract

Because the labor needed to manually label a huge training sample set is usually not available, the problem of hyperspectral image classification often suffers from a lack of labeled training samples. At the same time, hyperspectral data represented in a large number of bands are usually highly correlated. In this paper, to overcome the small sample problem in hyperspectral image classification, correlation of spectral bands is fully utilized to generate multiple new sub-samples from each original sample. The number of labeled training samples is thus increased several times. Experiment results demonstrate that the proposed method has an obvious advantage when the number of labeled samples is small.

Keywords

hyperspectral image / remote sensing / image classification / small sample problem

References

[1] Zhong Y, Zhang L. An adaptive artificial immune network for supervised classification of multi-hyperspectral remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing, 2011, 50(3): 894–909

[2] Pushmeet K, Lubor L, Philip H. Robust higher order potentials for enforcing label consistency. International Journal of Computer Vision, 2009, 82(3): 302–324

[3] Lubor L. Global structured models towards scene understanding. Dissertation for the Doctoral Degree. Oxford: Oxford Brooks Univerty, 2011

[4] Yang Y, Wu F, Nie F, Shen H, Zhuang Y, Alexander G. Web and personal image annotation by mining label correlation with relaxed visual graph embedding. IEEE Transactions on Image Processing, 2012, 21(3): 1339–1351

[5] Feng J, Jiao L, Zhang X, Sun T. Hyperspectral band selection based on trivariate mutual inforamtion and clonal selection. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(7): 4092–4105

[6] Qian D, He Y. Similarity-based unsupervised band selection for hyperspectral image analysis. IEEE Geoscience and Remote Sensing Letters, 2008, 5(4): 564–568

[7] Gustavo C, Joris M, Bernhard S. Remote sensing feature selection by kernel dependence measures. IEEE Geoscience and Remote Sensing Letters, 2010, 7(3): 587–591

[8] He Y, Qian D, Su H, Sheng Y. An efficient method for supervised hyperspectral band selection. IEEE Geoscience and Remote Sensing Letters, 2011, 8(1): 138–142

[9] Shen L, Zhu Z, Jia S, Zhu J, Sun Y. Discriminative gabor feature selection for hyperspectral image classification. IEEE Geoscience and Remote Sensing Letters, 2013, 10(1): 29–33

[10] Qian Y, Ye M, Zhou J. Hyperspectral image classification based on structured sparse logistic regression and three-dimensional wavelet texure features. IEEE Transactions on Geoscience and Remote Sensing, 2012, 51(4): 2276–2291

[11] Claude C, Kacem C, Steven L. BandClust: an unsupervised band reduction method for hyperspectral remote sensing. IEEE Geoscience and Remote Sening Letters, 2011, 8(3): 565–569

[12] Chris B. Pattern Recognition and Machine Learning. Springer, 2006,325–345

[13] Cai X, Wen G, Wei J, Yu Z. Relative manifold based semi-supervised dimensionality. Frontiers of Computer Science, 2014, 8(6): 923–932

[14] Li W, Guo Q, Charles E. A positive and unlabeled learning algorithm for one-class classification of remote-sensing data. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(2): 717–725

[15] Priscilla R, Swamynathan S. A semi-supervised hierarchical approach: two dimensional clustering of microarray gene expression data. Frontiers of Computer Science, 2013, 7(2): 204–213

[16] Yang L, Yang S, Jin P, Zhang R. Semis-upervised hyperspectral image classification using spatio-spectral laplacian support vector machine. IEEE Geoscience and Remote Sening Letters, 2014, 11(3): 651–655

[17] Wei D, Melba M. Active learning via multi-view and local proximity co-regulatization for hyperspectral image classification. IEEE Journal of Selected Topics in Signal Processing, 2011, 5(3): 618–628

[18] Swarnajyoti P, Lorenzo B. A fast cluster-assumption based activelearning technique for classification of remote sensing images. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(5): 1617–1626

[19] Begum D, Claudio P, Lorenzo B. Batch-mode active-learning methods for the interactive classification of remote sensing images. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(3): 1014–1031

[20] Gregg V, Robert O, Thomas G, Harry T, Earl G, Wallace M. The airboren visible/infrared imaging spectrometer. Remote Sensing of Enviroment, 1993, 44(3): 127–143

[21] Joseph L, Nicewander W. Thirteen ways to look at the correlation coefficient. The American Statistician, 1988, 42(1): 59–66

[22] Warner T, Steinmaus K, Foote B. An evaluation of spatial autocorrelation feature selection. International Jounral of Remote Sensing, 1999, 20(8): 1601–1606

[23] Masashi S. Dimensionality reduction of multimodal labeled data by local fisher discriminant analysis. Journal of Machine Learning Research, 2007, 8(3): 1027–1061

[24] Landgrebe D. Signal Theory Methods in Multispectral Remote Sensing. Hoboken: Wiley, 2003

[25] Yu X, Yang J, Xie Z. Training SVMs on a bound vectors set based on fisher projection. Frontiers of Computer Science, 2014, 8(5): 793–806

RIGHTS & PERMISSIONS

2014 Higher Education Press and Springer-Verlag Berlin Heidelberg