In this paper we examine the efficacy of using the closest distance to center algorithm in conjunction with ellipsoidal multivariate trimming (MVT) to find outliers in a hyperspectral image. MVT is applied here as a global anomaly detector on images that are pre-processed into clusters using a technique called X-means. Under the assumption that there are no more than 5% outliers in any given cluster set, we develop a method, based upon principal component analysis pre-processing, to create a flexible threshold for determining the percentage of data to retain with MVT. Using a retention percentage that more adequately reflects the actual number of outlier-free observations allows one to form estimates of the mean and covariance matrix that more effectively decrease the effects of swamping and masking as compared to using a set percentile for retention. These ideas are tested against real and synthetically generated hyperspectral imagery.
ArdouinJ-PLevesqueJReaTA. A demonstration of hyperspectral image exploitation for military applications. In: Proceedings of the 10th Conference on Information Fusion, Quebec, 9–12 July 2007 pp.1–8.
3.
BriottetXBoucherYDimmelerAMalaplateACiniADianiM. Military applications of hyperspectral imagery, targets and backgrounds XII: characterization and representation. Proc SPIE2006; 6239: 62390B.
4.
SteinDWJBeavenSGHoffLEWinterEMSchaumAPStockerAD. Anomaly detection from hyperspectral imagery. IEEE Signal Process Mag2002; 19: 58–69.
5.
ReedISYuX. Adaptive multiple-band CFAR detection of an optical pattern with unknown spectral distribution. IEEE Trans Acoust Speech Signal Process1990; 38: 1760–1770.
6.
YuXHoffLEReedISChenAMStottsLB. Automatic target detection and recognition in multiband imagery: a unified ML detection and estimation approach. IEEE Trans Image Process1997; 6: 143–156.
7.
TaitanoYPGeierBABauerKWJr. A locally adaptable iterative RX detector. EURASIP J Adv Signal Process2010; Article ID 341908.
8.
WilsonJR. Antithetic sampling with multivariate inputs. Am J Math Manage Sci1983; 3: 121–144.
9.
ChiangLHPellRJSeasholtzMB. Exploring process data with the use of robust outlier detection algorithms. J Process Contr2003; 13: 437–449.
10.
EganWJMorganSL. Outlier detection in multivariate analytical chemical data. Anal Chem1998; 77: 2372–2379.
11.
LandgrebeDA. Signal theory methods in multispectral remote sensing. Hoboken, NJ: John Wiley & Sons, 2003.
12.
RichardsJAJiaX. Remote Sensing Digital Image Analysis: An Introduction. Berlin: Springer, 1999, p. 363.
13.
BruceLMKogerCHLiJ. Dimensionality reduction of hyperspectral data using discrete wavelet transform feature extraction. IEEE Trans Geosci Remote Sens2002; 40: 2331–2338.
14.
KaewpijitSLe MoigneJEl-GhazawiT. Automatic reduction of hyperspectral imagery using wavelet spectral analysis. IEEE Trans Geosci Remote Sens2003; 41:863–871.
15.
DillonWRGoldsteinM. Multivariate analysis: methods and applications. New York: John Wiley & Sons, Inc., 1984.
16.
GnanadesikanRKettenringJR. Robust estimates, residuals, and outlier detection with multiresponse data. Biometrics1972; 28: 81–124.
17.
RaoCR. The use and interpretation of principal component analysis in applied research. Sankhya A1964; 26: 329–358.
18.
WilliamsJP. Robustness of multiple clustering algorithms on hyperspectral images. MS thesis, School of Operation Sciences, Air Force Institute of Technology (AU), Wright-Patterson AFB, OH, March2007.
19.
PellegDMooreA. X-means: extended K-means with efficient estimation of the number of clusters. In: Proceedings of the 17th International Conference on Machine Learning, 2000, pp.727–734.
20.
DevlinSJGnandesikanRKettenringJR. Robust estimation of dispersion matrices and principal components. J Am Stat Assoc1981; 76: 354–362.
21.
CaulkRF. Outlier detection in hyperspectral imagery using closest distance to center with ellipsoidal multivariate trimming. MS thesis, School of Operation Sciences, Air Force Institute of Technology (AU), Wright-Patterson AFB, OH, March2007.
22.
BillorNHadiASVellemanPF. BACON: blocked adaptive computationally efficient outlier nominators. Comput Stat Data Anal2000; 34: 279–298.
23.
SmetekTEBauerKW. A comparison of multivariate outlier detection methods for finding hyperspectral anomalies. Mil Oper Res2008; 13: 19–44.
24.
BellucciJPSmetekTEBauerKW. Improved hyperspectral image processing algorithm testing using synthetic imagery and factorial designed experiments. IEEE Trans Geosci Remote Sens2010; 48: 1211–1223.
