Abstract
The objective of this paper is to develop an algorithm to detect anomaly in a hyperspectral image. The algorithm is based on a subspace model that is derived statistically. The anomaly detector is defined as the Mahalanobis distance of a residual from a pixel that is partitioned uniformly. The high correlation among adjacent components of the pixel is exploited by partitioning the pixel uniformly to improve anomaly detection. The residual is obtained by partialling out the main background from the pixel by predicting a linear combination of each partition of the pixel with a linear combination of the random variables representing the main background. Experimental results show that the anomaly detector outperforms conventional anomaly detectors.
Similar content being viewed by others
References
Bioucas-Dias JM, Plaza A, Camps-Valls G, Scheunders P, Nasrabadi NM, Chanussot J (2013) Hyperspectral remote sensing data analysis and future challenges. IEEE Geosci Remote Sens Mag 1(2):6–36
Matteoli S, Diani M, Corsini G (2010) A tutorial overview of anomaly detection in hyperspectral images. IEEE Trans Aerosp Electron Systems Mag 25(7):5–27
Stein DWJ, Beaven SG, Hoff LE, Winter EM, Schaum AP, Stocker AD (2002) Anomaly detection from hyperspectral imagery. IEEE Signal Process Mag 19(1):58–69
Chen JY, Reed IS (1987) A detection algorithm for optical targets in clutter. IEEE Trans Aerosp Electron Systems 23(1):394–405
Stocker A, Schaum A (1997) Application of stochastic mixing models to the hyperspectral detection problem. Proc SPIE 3071:47–60
Kwon H, Der SZ, Nasrabadi NM (2003) Using self-organizing maps for anomaly detection in hyperspectral imagery. Optical Eng 42(11):3342–3351
Kwon H, Nasrabadi NM (2005) Kernel RX-algorithm: a nonlinear anomaly detector for hyperspectral imagery. IEEE Trans Geosci Remote Sens 43(2):388–397
Reed IS, Yu X (1990) Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution. IEEE Trans Acoust Speech Signal Process 38(10):1760–1770
Lo E, Ingram J (2008) Hyperspectral anomaly detection based on minimum generalized variance method. Proc SPIE, vol. 6966, no. 696603
Hastie T, Tibshirani R, Friedman J (2001) The elements of statistical learning. Springer, Berlin
Bachega LR, Theiler J (2011) Evaluating and improving local hyperspectral anomaly detectors. IEEE Applied Imagery Pattern Recognition Workshop (AIPR), pp. 1–8
Chang CI (2005) Orthogonal subspace projection (OSP) revisited: a comprehensive study and analysis. IEEE Trans Geosci Remote Sens 43(3):502–518
Ranney KI, Soumekh M (2006) Hyperspectral anomaly detection within the signal subspace. IEEE Geosci Remote Sens Lett 3(3):312–316
Duran O, Petrou M (2009) Spectral unmixing with negative and superunity abundances for subpixel anomaly detection. IEEE Trans Geosci Remote Sens Lett 6(1):152–156
Goldberg H, Nasrabadi N (2007) A comparative study of linear and nonlinear anomaly detectors for hyperspectral imagery. Proceedings of SPIE, vol. 6497
Goldberg H, Kwon H, Nasrabadi NM (2007) Kernel eigenspace separation transform for subspace anomaly detection in hyperspectral imagery. IEEE Geosci Remote Sens Lett 4(4):581–585
Schaum AP (2004) Joint subspace detection of hyperspectral targets. Proceeding of IEEE Aerospace Conference, vol. 3
Schaum AP (2007) Hyperspectral anomaly detection beyond RX. Proceeding of SPIE, vol. 6565, no. 656502
Lo E (2014) Variable factorization model based on numerical optimization for hyperspectral anomaly detection. Pattern Anal Appl 17(2):291–310
Lo E (2012) Maximized subspace model for hyperspectral anomaly detection. Pattern Anal Appl 15(3):225–235
Lo E (2013) Variable subspace model for hyperspectral anomaly detection. Pattern Anal Appl 16(3):393–405
Lo E, Ingram J (2011) Algorithm for detecting anomaly in hyperspectral imagery using factor analysis. Proceedings of SPIE, vol. 8048, no. 804805
Kerekes JP, Snyder DK (2010) Unresolved target detection blind test project overview. Proceeding of 16th SPIE Conference on algorithms and technologies for multispectral, hyperspectral, and ultraspectral imagery, vol. 7695, no. 769521
Acknowledgments
The author wishes to thank the U.S. Naval Research Laboratory and the Rochester Institute of Technology for the data.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Lo, E. Hyperspectral anomaly detection based on uniformly partitioned pixel. Pattern Anal Applic 19, 297–309 (2016). https://doi.org/10.1007/s10044-014-0398-2
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10044-014-0398-2