Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Spectroscopy and Spectral Analysis, Volume. 42, Issue 4, 1270(2022)

Hyperspectral Anomaly Detection Based on 3D Convolutional Autoencoder Network

Sheng-ming WANG1、*, Tao WANG1、1; *;, Sheng-jin TANG2、2;, and Yan-zhao SU1、1;
Author Affiliations
  • 11. Combat Support Academy, Rocket Force University of Engineering, Xi’an 710025, China
  • 22. Missile Engineering Academy, Rocket Force University of Engineering, Xi’an 710025, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (12)
    Equations (0)
    References (13)
    Tools
    Paper Information
    Topics
    Figures & Tables(12)
    Comparing 2D convolution operation (a) and 3D convolution operation (b)

    Fig. 1. Comparing 2D convolution operation (a) and 3D convolution operation (b)

    Download full size
    View in Article
    3D convolution

    Fig. 2. 3D convolution

    Download full size
    View in Article
    Unsupervised anomaly detection framework based on 3D-CAE

    Fig. 3. Unsupervised anomaly detection framework based on 3D-CAE

    Download full size
    View in Article
    Results of San Diego datasets anomaly detection(a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Fig. 4. Results of San Diego datasets anomaly detection
    (a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Download full size
    View in Article
    Results of Los Angeles datasets anomaly detection(a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Fig. 5. Results of Los Angeles datasets anomaly detection
    (a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Download full size
    View in Article
    Results of Pavia datasets anomaly detection(a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Fig. 6. Results of Pavia datasets anomaly detection
    (a): False color image; (b): Reference map; (c): RX; (d): SRX; (e): CRD; (f): UNRS; (g): LRASR; (h): 3D-CAEAD

    Download full size
    View in Article
    ROC curves of San Diego datasets

    Fig. 7. ROC curves of San Diego datasets

    Download full size
    View in Article
    ROC curves of Los Angeles datasets

    Fig. 8. ROC curves of Los Angeles datasets

    Download full size
    View in Article
    ROC curves of Pavia datasets

    Fig. 9. ROC curves of Pavia datasets

    Download full size
    View in Article

    • Table 1. Information of datasets

      View table
      View in Article

      Table 1. Information of datasets

      名称传感器波长/μm像素波段
      San Diego-1AVRIS0.37~2.51100×100205
      San Diego-2AVRIS0.37~2.51100×100191
      San Diego-3AVRIS0.37~2.51120×120189
      San Diego-4AVRIS0.37~2.51100×100189
      Los Angeles-1AVRIS0.45~1.35100×100204
      Los Angeles-2AVRIS0.45~1.35100×100207
      Pavia-1ROSIS0.43~0.86100×100188
      Pavia-2ROSIS0.43~0.86100×100188
      Pavia-3ROSIS0.43~0.86100×100102

    • Table 2. 3D-CAE parameter settings

      View table
      View in Article

      Table 2. 3D-CAE parameter settings

      卷积层INPUT
      SIZE      		KERNEL
      SIZE      		KERNEL
      NUMBERS      		STRIDEPADDINGOUTPUT
      SIZE
      CONV1224 5 5 12×3×3322 1 1valid112 3 3 32
      CONV2112 3 3 322×1×1642 1 1valid56 3 3 64
      CONV356 3 3 642×3×31282 1 1valid28 1 1 128
      DECONV128 1 1 1282×3×3642 1 1valid56 3 3 64
      DECONV256 3 3 642×1×1322 1 1valid112 3 3 32
      DECONV3112 3 3 322×3×312 1 1valid224 5 5 1

    • Table 3. AUC values of three groups of datasets

      View table
      View in Article

      Table 3. AUC values of three groups of datasets

      数据集RXSRXCRDUNRSLRASR3D-CAEAD
      San Diego-10.840 40.902 50.884 00.898 70.866 00.971 7
      San Diego-20.952 60.944 10.768 10.791 70.962 20.986 9
      San Diego-30.911 10.949 50.760 20.812 60.988 70.978 4
      San Diego-40.940 30.961 70.907 10.919 80.894 90.983 3
      Los Angeles-10.990 70.724 10.965 50.972 20.954 00.993 0
      Los Angeles-20.944 00.927 50.922 40.981 60.989 90.984 6
      Pavia-10.980 70.966 40.992 80.992 90.918 70.987 8
      Pavia-20.999 10.995 80.994 40.997 60.996 30.999 9
      Pavia-30.953 80.877 70.886 80.959 90.930 50.972 4
    Tools

    Get Citation

    Copy Citation Text

    Sheng-ming WANG, Tao WANG, Sheng-jin TANG, Yan-zhao SU. Hyperspectral Anomaly Detection Based on 3D Convolutional Autoencoder Network[J]. Spectroscopy and Spectral Analysis, 2022, 42(4): 1270

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Mar. 13, 2021

    Accepted: --

    Published Online: Jul. 25, 2023

    The Author Email: WANG Sheng-ming (210279598@qq.com)

    DOI:10.3964/j.issn.1000-0593(2022)04-1270-08

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology