MRF Based Spatial Complexity for Hyperspectral Imagery Unmixing
Hyperspectral imagery (HSI) unmixing is a process that decomposes pixel spectra into a collection of constituent spectra (endmembers) and their correspondent abundance fractions. Without knowing any knowledge of HSI data, the unmixing problem is transformed into a blind source separation (BSS) problem. Several methods have been proposed to deal with the problem, like independent component analysis (ICA). In this paper, we introduce spatial complexity that applies Markov random field (MRF) to characterize the spatial correlation information of abundance fractions. Compared to previous BSS techniques for HSI unmixing, the major advantage of our approach is that it totally considers HSI spatial structure. Additionally, a proof is given that spatial complexity is suitable for HSI unmixing. Encouraging results have been obtained in terms of unmixing accuracy, suggesting the effectiveness of our approach.
KeywordsIndependent Component Analysis Hyperspectral Image Markov Random Field Independent Component Analysis Blind Source Separation
Unable to display preview. Download preview PDF.
- 2.Keshava, N.: A survey of spectral unmixing algorithms. Lincoln Lab Journal 14(1), 55–73 (2003)Google Scholar
- 3.Yuhas, R.H., Goetz, A.F.H., Boardman, J.W.: Discrimination among semi-arid landscape endmembers using the spectral angle mapper (sam) algorithm. In: Summaries of the 3rd Annual JPL Airborne Geoscience Workshop, vol. 1, pp. 147–149 (1992)Google Scholar
- 5.Bayliss, J.D., Gualtieri, J.A., Cromp, R.F.: Analyzing hyperspectral data with independent component analysis. In: Proceeding of SPIE Applied Image and Pattern Recognition Workshop, vol. 3240, pp. 133–143 (1997)Google Scholar
- 8.Stone, J.V.: Independent Component Analysis: A Tutorial Introduction. MIT Press, Cambridge (2004)Google Scholar
- 9.Du, Q., Chakrarvarty, S.: Unsupervised hyperspectral image classification using blind source separation. In: IEEE International Conference on Acoustics, Speech, and Signal Processing 2003, vol. 3, pp. 437–440 (2003)Google Scholar
- 12.Robila, S.A., Varshney, P.K.: Target detection in hyperspectral images based on independent component analysis. In: Proceeding of SPIE Automatic Target Recognition, vol. 4726, pp. 173–182 (2002)Google Scholar
- 14.The HYDICE HSI dataset, http://www.tec.army.mil/Hypercube/
- 16.Stone, J.V., Porrill, J.: Undercomplete independent component analysis for signal separation and dimension reduction. Technical report, Psychology Department, Sheffield University (1998), http://www.shef.ac.uk/~pc1jvs/