A Discriminative Metric Learning Based Anomaly Detection Method

被引:239
作者
Du, Bo [1 ]
Zhang, Liangpei [2 ]
机构
[1] Wuhan Univ, Sch Comp Sci, Wuhan 430072, Peoples R China
[2] Wuhan Univ, Remote Sensing Grp, State Key Lab Informat Engn Surveying Mapping & R, Wuhan 430072, Peoples R China
来源
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING | 2014年 / 52卷 / 11期
基金
中国国家自然科学基金;
关键词
Anomaly detection; hyperspectral images; image processing; NONLINEAR DIMENSIONALITY REDUCTION; TARGET DETECTION; CLASSIFICATION; ALGORITHM;
D O I
10.1109/TGRS.2014.2303895
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
Due to the high spectral resolution, anomaly detection from hyperspectral images provides a new way to locate potential targets in a scene, especially those targets that are spectrally different from the majority of the data set. Conventional Mahalanobis-distance-based anomaly detection methods depend on the background statistics to construct the anomaly detection metric. One of the main problems with these methods is that the Gaussian distribution assumption of the background may not be reasonable. Furthermore, these methods are also susceptible to contamination of the conventional background covariance matrix by anomaly pixels. This paper proposes a new anomaly detection method by effectively exploiting a robust anomaly degree metric for increasing the separability between anomaly pixels and other background pixels, using discriminative information. First, the manifold feature is used so as to divide the pixels into the potential anomaly part and the potential background part. This procedure is called discriminative information learning. A metric learning method is then performed to obtain the robust anomaly degree measurements. Experiments with three hyperspectral data sets reveal that the proposed method outperforms other current anomaly detection methods. The sensitivity of the method to several important parameters is also investigated.
引用
收藏
页码:6844 / 6857
页数:14
相关论文
共 50 条
  • [1] INTERIOR-POINT METHODS IN SEMIDEFINITE PROGRAMMING WITH APPLICATIONS TO COMBINATORIAL OPTIMIZATION
    ALIZADEH, F
    [J]. SIAM JOURNAL ON OPTIMIZATION, 1995, 5 (01) : 13 - 51
  • [2] A support vector method for anomaly detection in hyperspectral imagery
    Banerjee, Amit
    Burlina, Philippe
    Diehl, Chris
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2006, 44 (08): : 2282 - 2291
  • [3] Bengio Y, 2004, ADV NEUR IN, V16, P177
  • [4] BACON: blocked adaptive computationally efficient outlier nominators
    Billor, N
    Hadi, AS
    Velleman, PF
    [J]. COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2000, 34 (03) : 279 - 298
  • [5] Improved covariance matrices for point target detection in hyperspectral data
    Caefer, Charlene E.
    Silverman, Jerry
    Orthal, Oded
    Antonelli, Dani
    Sharoni, Yaron
    Rotman, Stanley R.
    [J]. OPTICAL ENGINEERING, 2008, 47 (07)
  • [6] Semi-supervised graph-based hyperspectral image classification
    Camps-Valls, Gustavo
    Bandos, Tatyana V.
    Zhou, Dengyong
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2007, 45 (10): : 3044 - 3054
  • [7] A cluster-based approach for detecting man-made objects and changes in imagery
    Carlotto, MJ
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2005, 43 (02): : 374 - 387
  • [8] Chang C.I., 2003, Hyperspectral Imaging: Techniques for Spectral Detection and Classification, V1
  • [9] Anomaly detection and classification for hyperspectral imagery
    Chang, CI
    Chiang, SS
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2002, 40 (06): : 1314 - 1325
  • [10] Estimation of number of spectrally distinct signal sources in hyperspectral imagery
    Chang, CI
    Du, Q
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2004, 42 (03): : 608 - 619