Integration of spectral and spatial information via local covariance matrices for segmentation and classification of hyperspectral images

In this work, a novel approach is presented for the feature extraction step in hyperspectral image processing to form more discriminative features between different pixel regions. The proposed method combines both spatial and spectral information, which is very important for segmentation and classification of hyperspectral images. For comparison, five different feature sets are formed using eigen decomposition of local covariance matrices of subcubes located around a pixel of interest in the scene. Subcubes of neighbor pixels are obtained by a windowed structure to expose pattern similarities. As a novel approach, local covariance matrices are computed in eigenspace and proposed feature sets are created after this stage. Before the formation of feature sets in eigenspace, the original input space is transferred to eigenspace by linear and nonlinear manner by principal component analysis (PCA) and its kernelized version (KPCA) and they are used in the experiments comparatively. In the simulations, one hyperspectral scene with ground-truth and one without ground-truth are used for the segmentation and classification tasks. Results of experiments are evaluated with four different unsupervised learning algorithms for data without ground-truth and three different supervised learning algorithms for data with ground-truth comparatively.

PDF

___

[1] Landgrebe DA. Signal Theory Methods in Multispectral Remote Sensing. Hoboken, NJ, USA: Wiley, 2003.
[2] Varshney PK, Arora MK. Advanced Image Processing Techniques for Remotely Sensed Hyperspectral Data. Berlin, Germany: Springer, 2004.
[3] Plaza A, Benediktsson JA, Boardman JW, Brazile J, Bruzzone L, Camps-Valls G, Chanussot J, Fauvel M, Gamba P, Gualtieri A et al. Recent advances in techniques for hyperspectral image processing. Remote Sens Environ 2012; 113: 110-122.
[4] Meyer AW, Paglieroni DW, Astaneh C. K-means re-clustering: algorithmic options with quantifiable performance comparisons. In: International Society for Optics and Photonics High-Power Lasers and Applications Conference; 25 Jan 2003; San Jose, CA, USA. Bellingham, WA, USA: SPIE. pp. 84-92.
[5] Kuo BC, Huang WC, Liu HC, Tseng SC. A novel fuzzy c-means method for hyperspectral image classification. In: IEEE International Geoscience and Remote Sensing Symposium; 711 July 2008; Boston, MA, USA. Hoboken, NJ, USA: IEEE. pp. 1002-1005.
[6] Muhammed HH. Unsupervised hyperspectral image segmentation using a new class of neuro-fuzzy systems based on weighted incremental neural networks. In: The 31st Applied Imagery Pattern Recognition Workshop; 1617 October 2002; Washington DC, USA. Piscataway, NJ, USA: IEEE. pp. 171-177.
[7] Erturk A, Erturk S. Unsupervised segmentation of hyperspectral images using modified phase correlation. IEEE Geosci Remote S 2006; 3: 527-531.
[8] Paoli A, Melgani F, Pasolli E. Clustering of hyperspectral images based on multiobjective particle swarm optimization. IEEE T Geosci Remote 2009; 47: 4175-4188.
[9] Mercier G, Lennon M. Support vector machines for hyperspectral image classification with spectral-based kernels. In: IEEE International Geoscience and Remote Sensing Symposium; 2125 July 2003; Toulouse, France. Hoboken, NJ, USA: IEEE. pp. 288-290.
[10] Camps-Valls G, Bruzzone L. Kernel Methods for Remote Sensing Data Analysis. New York, NY, USA: Wiley Online Library, 2009.
[11] Demir B, Erturk S. Hyperspectral image classification using relevance vector machines. IEEE Geosci Remote S 2007; 4: 586-590.
[12] Camps-Valls G, Gomez-Chova L, Munoz-Mari J, Vila-Frances J, Calpe-Maravilla J. Composite kernels for hyperspectral image classification. IEEE Geosci Remote S 2006; 3: 93-97.
[13] Chen Y, Nasrabadi NM, Tran TD. Hyperspectral image classification via kernel sparse representation. IEEE T Geosci Remote S 2013; 51: 217-231.
[14] Camps-Valls G, Marsheva TB, Zhou D. Semi-supervised graph-based hyperspectral image classification. IEEE T Geosci Remote S 2007; 45: 3044-3054.
[15] Ratle F, Camps-Valls G, Weston J. Semisupervised neural networks for efficient hyperspectral image classification. IEEE T Geosci Remote S 2010; 48: 2271-2282.
[16] Ma L, Crawford MM, Tian J. Local manifold learning-based-nearest-neighbor for hyperspectral image classification. IEEE T Geosci Remote S 2010; 48: 4099-4109.
[17] Wang C, Gao RX. Wavelet transform with spectral post-processing for enhanced feature extraction-machine condition monitoring. IEEE Instru Meas Mag 2003; 52: 1296-1301.
[18] Bruce LM, Koger CH, Li J. Dimensionality reduction of hyperspectral data using discrete wavelet transform feature extraction. IEEE T Geosci Remote 2002; 40: 2331-2338.
[19] Mendenhall MJ, Mer´enyi E. Relevance-based feature extraction for hyperspectral images. IEEE T Neural Networ 2008; 19: 658-672.
[20] Zhang L, Zhang L, Tao D, Huang X. Tensor discriminative locality alignment for hyperspectral image spectral spatial feature extraction. IEEE T Geosci Remote 2013; 51: 242-256.
[21] Mojaradi B, Abrishami-Moghaddam H, Zoej MJV, Duin RP. Dimensionality reduction of hyperspectral data via spectral feature extraction. IEEE T Geosci Remote 2009; 47: 2091-2105.
[22] Fauvel M, Benediktsson JA, Chanussot J, Sveinsson JR. Spectral and spatial classification of hyperspectral data using SVMs and morphological profiles. IEEE T Geosci Remote 2008; 46: 3804-3814.
[23] DellAcqua F, Gamba P, Ferrari A, Palmason JA, Benediktsson JA, Arnason KAAK. Exploiting spectral and spatial information in hyperspectral urban data with high resolution. IEEE Geosci Remote S 2004; 1: 322-326.
[24] Tarabalka Y, Benediktsson JA, Chanussot J. Spectralspatial classification of hyperspectral imagery based on partitional clustering techniques. IEEE T Geosci Remote 2009; 47: 2973-2987.
[25] Mar¸cal ARS, Castro L. Hierarchical clustering of multispectral images using combined spectral and spatial criteria. IEEE Geosci Remote S 2005; 2: 59-63.
[26] Li J, Bioucas-Dias JM, Plaza A. Spectralspatial hyperspectral image segmentation using subspace multinomial logistic regression and Markov random fields. IEEE T Geosci Remote 2012; 50: 809-823.
[27] Huo LZ, Tang P. Spectral and spatial classification of hyperspectral data using SVMs and Gabor textures. In: IEEE International Geoscience and Remote Sensing Symposium; 2429 July 2011; Vancouver, Canada. Hoboken, NJ, USA: IEEE. pp. 1708-1711.
[28] Bilgin G, Uslu E. Segmentation of hyperspectral images using local covariance matrices. In: 20th Signal Processing and Communications Applications Conference; 1820 April 2012; Fethiye, Turkey. Piscataway, NJ, USA: IEEE. pp. 1-4.
[29] Ergul U, Bilgin G. Segmentation of hyperspectral images using local covariance matrices in eigenspace. In: 21st Signal Processing and Communications Applications Conference; 2426 April 2013; Girne, Cyprus. Piscataway, NJ, USA: IEEE. pp. 1-4.
[30] Van der Meer F. The effectiveness of spectral similarity measures for the analysis of hyperspectral imagery. Int J Appl Earth Obs 2006; 8: 3-17.
[31] Haralick RM, Shanmugam K, Dinstein IH. Textural features for image classification. IEEE T Syst Man Cyb 1973; 3: 610-621.
[32] Huang X, Zhang L. A comparative study of spatial approaches for urban mapping using hyperspectral ROSIS images over Pavia City, Northern Italy. Int J Remote Sens 2009; 30: 3205-3221.
[33] Jimenez-Rodriguez OL, Arzuaga-Cruz E, Velez-Reyes M. Unsupervised linear feature-extraction methods and their effects in the classification of high-dimensional data. IEEE T Geosci Remote 2007; 45: 469-483.
[34] Mathieu F, Chanussot J, Benediktsson JA. Kernel principal component analysis for the classification of hyperspectral remote sensing data over urban areas. EURASIP J Adv Sig Pr 2009; 11: 1-14.
[35] Tuzel O, Porikli F, Meer P. Region covariance: A fast descriptor for detection and classification. In: European Conference on Computer Vision; 713 May 2006; Graz, Austria. Berlin, Germany: Springer. pp. 589-600.
[36] Swayze G, Clark R, Kruse F, Sutley S, Gallagher A. Groundtruthing AVIRIS mineral mapping at Cuprite, Nevada. In: Summaries of the Third Annual JPL Airborne Geoscience Workshop; 1 Jun 1992; Denver, CO, USA. pp. 47-49.
[37] Chi M, Kun Q, Benediktsson JA, Feng R. Ensemble classification algorithm for hyperspectral remote sensing data. IEEE T Geosci Remote 2009; 6: 762-766.