Optics and Precision Engineering, Volume. 32, Issue 15, 2439(2024)

Visible-polarized image fusion for nighttime dispersal of mines

Kangyu SHEN1... Bolun CUI2, Qifeng LYU1 and Chi WANG1,* |Show fewer author(s)
Author Affiliations
  • 1School of Mechanical Engineering and Automation, Shanghai University, Shanghai200444, China
  • 2Beijing Institute of Space Mechanics & Electricity, Beijing100094, China
  • show less
    Figures & Tables(17)
    Network model of visible-polarized image fusion algorithm
    Structure of Sobel Gradient Dense Block (SGDB)
    Structure of attention mechanism model
    Targets for mines used in dataset
    Multi-angle polarization observation devices
    Low-light camera
    Targets for mines used in datasets
    Fused images of different methods in sandy soil background
    Fused images from different methods in complex environment with multiple deciduous leaves
    Fused images with different equilibrium coefficients α
    Fused images generated by models with different attention mechanisms
    Effect of visible intensity images and fused images on target detection performance
    Detection effect of visible intensity images and fused images in part of the test set
    • Table 1. Objective evaluation results of fused images by different methods

      View table
      View in Article

      Table 1. Objective evaluation results of fused images by different methods

      MethodsSSIMMS-SSIMPSNRMSEENSDAGSCDVIFQabfMISF
      Our method0.758 50.781 327.8453.215.70719.683.8191.5970.837 30.606 52.68910.69
      CVT0.737 40.779 325.9768.195.20513.443.8031.5360.709 60.545 81.44410.46
      NSCT0.742 40.780 726.0167.735.20313.543.7191.5710.778 90.584 61.54910.50
      SR0.723 70.739 228.9246.265.55619.563.3371.3600.710 90.473 02.80110.05
      NSCT-ST0.731 20.742 628.9646.215.60519.573.5931.4110.739 20.512 62.81110.49
      FusionGAN0.608 00.702 324.5191.865.54417.922.3271.1750.500 60.205 31.3755.499
      DenseFusion0.677 20.746 324.9886.825.70321.953.8041.4610.819 30.485 22.62311.14
      PFNet0.746 80.770 527.3857.555.38119.493.7111.5010.746 50.565 32.54414.55
    • Table 2. Effect of balancing coefficient α on quality of fused images

      View table
      View in Article

      Table 2. Effect of balancing coefficient α on quality of fused images

      αSSIMMS-SSIMPSNRMSEENSDAGSCDVIFQabfMISF
      00.755 60.784 027.1852.675.31718.443.3461.5730.799 80.545 73.0829.868
      0.50.746 60.775 227.8648.625.50319.613.8541.6100.910 10.645 03.00410.71
      10.745 10.775 427.9947.505.51219.493.8581.6400.910 40.645 72.94810.72
      50.743 90.775 327.6050.595.49519.043.8381.6680.913 80.646 63.06110.71
      100.739 70.774 526.7857.955.46519.033.8291.6600.909 50.646 73.00310.68
      1000.083 90.477 623.61109.31.8102.4700.8120.410 40.030 10.066 30.493 22.579
      10000.608 20.644 524.01101.95.33417.443.9560.532 70.758 20.631 82.71211.48
    • Table 3. Effect of different balance coefficients λ on quality of fused images

      View table
      View in Article

      Table 3. Effect of different balance coefficients λ on quality of fused images

      λSSIMMS-SSIMPSNRMSEENSDAGSCDVIFQabfMISF
      00.745 10.775 227.9947.505.51219.493.8581.6400.910 40.644 72.94810.71
      0.10.758 50.779 927.8253.555.44819.673.8121.5920.844 40.600 32.68910.72
      0.50.759 50.781 327.6056.745.37519.193.7701.6110.819 00.585 12.56110.63
      10.759 10.779 127.7556.145.38919.663.7581.5540.813 10.576 72.51510.64
      100.760 40.781 527.7057.415.32918.873.7121.5740.793 10.568 72.44010.56
      1000.761 60.783 127.5458.345.30218.673.6951.5860.788 40.565 92.43110.51
      1 0000.761 10.781 527.7357.325.33519.093.7071.5740.791 90.566 42.44410.54
    • Table 4. Objective evaluation of model-generated fused images under different attention mechanisms

      View table
      View in Article

      Table 4. Objective evaluation of model-generated fused images under different attention mechanisms

      Fusion moduleSSIMMS-SSIMPSNRMSEENSDAGSCDVIFQabfMISF
      Lack of attention0.516 20.738 624.4395.863.1044.1540.662 40.965 20.181 00.025 11.3922.067
      Spatial attention0.752 40.798 625.8570.374.86311.472.736 01.387 00.507 10.378 01.9788.059
      Channel attention0.754 40.776 627.2557.535.40818.833.819 01.553 00.842 80.597 82.60810.66
      Mixed attention0.758 50.779 927.8253.555.44819.673.812 01.592 00.844 30.600 32.68910.71
    Tools

    Get Citation

    Copy Citation Text

    Kangyu SHEN, Bolun CUI, Qifeng LYU, Chi WANG. Visible-polarized image fusion for nighttime dispersal of mines[J]. Optics and Precision Engineering, 2024, 32(15): 2439

    Download Citation

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

    Category:

    Received: May. 17, 2024

    Accepted: --

    Published Online: Sep. 27, 2024

    The Author Email: WANG Chi (wangchi@shu.edu.cn)

    DOI:10.37188/OPE.20243215.2439

    Topics