• Laser & Optoelectronics Progress
  • Vol. 59, Issue 4, 0415005 (2022)
Yanling Du1, Jianhua Cui1, Quanmiao Wei2, and Dongmei Huang3、*
Author Affiliations
  • 1College of Information Technology, Shanghai Ocean University, Shanghai 201306, China
  • 2East China Sea Bureau of Ministry of Natural Resources, Shanghai 200136, China
  • 3Shanghai University of Electric Power, Shanghai 200090, China
  • show less
    DOI: 10.3788/LOP202259.0415005 Cite this Article Set citation alerts
    Yanling Du, Jianhua Cui, Quanmiao Wei, Dongmei Huang. Marine Oil-Spill Detection in Multi-Polarization Image-Based SAR on Improved FCN[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0415005 Copy Citation Text show less
    Intelligent detection framework of Pol-SAR image based on improved FCN
    Fig. 1. Intelligent detection framework of Pol-SAR image based on improved FCN
    Comparison of oil spill area detection under different polarization modes
    Fig. 2. Comparison of oil spill area detection under different polarization modes
    Comparison of oil spill detections of two polarimetric fusion SAR images. (a) HH-VV; (b) HH-HV; (c) HH-VH; (d) HV-VH; (e) HV-VV; (f) VH-VV
    Fig. 3. Comparison of oil spill detections of two polarimetric fusion SAR images. (a) HH-VV; (b) HH-HV; (c) HH-VH; (d) HV-VH; (e) HV-VV; (f) VH-VV
    Oil spill results of three-polarimetric fusion and full-polarimetric fusion. (a) HH-HV-VH; (b) HH-HV-VV; (c) HH-VH-VV; (d) HV-VH-VV; (e) HH-VV-HV-VH
    Fig. 4. Oil spill results of three-polarimetric fusion and full-polarimetric fusion. (a) HH-HV-VH; (b) HH-HV-VV; (c) HH-VH-VV; (d) HV-VH-VV; (e) HH-VV-HV-VH
    Results of oil spill area detected by different methods. (a) Original imags; (b) labeled imags; (c) FCN; (d) UNet; (e) Improved FCN
    Fig. 5. Results of oil spill area detected by different methods. (a) Original imags; (b) labeled imags; (c) FCN; (d) UNet; (e) Improved FCN
    ParameterSingle look complex(SLC)
    Start time2011-06-17T11∶48∶20
    Beam modeFQ25
    PolarizationHH,VV,HV,VH
    Look directionRight
    Pixel spacing /(m×m)4.73×5.05
    Incidence angle4.36E1-4.49E1
    Covered area /(km×km)37.17×19.34
    Table 1. Information on the RadarSAT-2 SAR image used in this study
    PolarizationIndexOriginal imagePauli decompositionPauli decomposition +Refined-Lee (3×3)Pauli decomposition +Refined-Lee (5×5)Pauli decomposition +Refined-Lee (7×7)
    VVPA0.9170.9270.9340.9210.920
    MA0.8960.9010.9220.9160.913
    HHPA0.9050.9160.9270.9230.919
    MA0.8680.9050.9110.9060.898
    Table 2. Detection result comparison of oil spill after Pol-SAR image preprocessing
    IndexHH-VVHH-HVHH-VHHV-VHHV-VVVH-VV
    PA0.9520.9390.9420.5050.9220.951
    MA0.9270.8930.9160.5160.9310.917
    Table 3. Comparison of oil spill detection accuracy of two polarimetric fusion SAR images
    IndexHH-HV-VHHH-HV-VVHH-VH-VVHV-VH-VVHH-VV-HV-VH
    PA0.9420.9130.9450.9490.957
    MA0.9060.9220.9190.8900.930
    Table 4. Comparison of oil spill detection accuracy of three-polarimetric fusion and full-polarimetric fusion
    Index(a)(b)(c)(d)
    PA0.9530.9470.9410.950
    MA0.9260.9220.9140.921
    Table 5. Accuracy of fully polarized oil spill detection in different regions
    AreaEvaluation indexFCNUNetImproved FCN
    Area 1PA0.8910.9130.948
    MA0.8340.8650.927
    Area 2PA0.8660.9050.953
    MA0.8520.8850.934
    Area 3PA0.8840.9160.941
    MA0.8350.8980.935
    Area 4PA0.8630.9270.936
    MA0.8150.9050.911
    Table 6. Oil spill detection accuracy in different regions
    Yanling Du, Jianhua Cui, Quanmiao Wei, Dongmei Huang. Marine Oil-Spill Detection in Multi-Polarization Image-Based SAR on Improved FCN[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0415005
    Download Citation