• Acta Photonica Sinica
  • Vol. 51, Issue 4, 0404003 (2022)
Sufeng ZHUANG, Yong JI, Dawei TU*, and Xu ZHANG
Author Affiliations
  • School of Mechatronic Engineering and Automation,Shanghai University,Shanghai 200444,China
  • show less
    DOI: 10.3788/gzxb20225104.0404003 Cite this Article
    Sufeng ZHUANG, Yong JI, Dawei TU, Xu ZHANG. Underwater RGB-D Camera Based on Binocular Stereo Vision[J]. Acta Photonica Sinica, 2022, 51(4): 0404003 Copy Citation Text show less
    Underwater camera imaging model
    Fig. 1. Underwater camera imaging model
    System Schematic
    Fig. 2. System Schematic
    Underwater binocular vision model
    Fig. 3. Underwater binocular vision model
    Pixel-based matching process diagram
    Fig. 4. Pixel-based matching process diagram
    Flow chart of color correction algorithm
    Fig. 5. Flow chart of color correction algorithm
    Enhanced effects of images
    Fig. 6. Enhanced effects of images
    Flow chart of alignment and superposition of 3D point cloud and RGB color
    Fig. 7. Flow chart of alignment and superposition of 3D point cloud and RGB color
    Alignment superposition model of 3D point cloud and 2D RGB data
    Fig. 8. Alignment superposition model of 3D point cloud and 2D RGB data
    System electrical structure diagram
    Fig. 9. System electrical structure diagram
    Physical engineering prototype
    Fig. 10. Physical engineering prototype
    Internal structure of the central cylinder
    Fig. 11. Internal structure of the central cylinder
    Internal structure of cylinder on both sides
    Fig. 12. Internal structure of cylinder on both sides
    Experimental scene
    Fig. 13. Experimental scene
    Original underwater images and enhanced images
    Fig. 14. Original underwater images and enhanced images
    Effect comparison of enhancement algorithms
    Fig. 15. Effect comparison of enhancement algorithms
    Registration effect of 3D point cloud and RGB color
    Fig. 16. Registration effect of 3D point cloud and RGB color
    Precision measurement experiment
    Fig. 17. Precision measurement experiment
    The imageThe evaluation criterionThe original imageHEIFDCPThe proposed
    1Information entropy6.837.287.247.147.47
    Sift1541 8489851 3612 079
    Canny18 66667 71453 25960 47167 963
    2Information entropy7.707.397.457.277.88
    Sift5821 5479619162 004
    Canny27 39267 93456 94332 46368 557
    3Information entropy7.157.237.327.197.70
    Sift5 95413 4888 62510 07217 203
    Canny130 335211 820167 198150 484211 804
    Table 1. the comparison result of different algorithms
    ParametersLeft cameraRight camera
    fx,fy(6 801.773,6 801.983)(6 766.651,6 767.138)
    u0,v0(2 032.804,1 456.358)(2 146.771,1 495.447)
    (k1,k2,k3,p1,p2)(-0.028,0.132,0.001,0.001,-0.465)(-0.027,0.137,0.001,-0.001,-0.485)
    d60.682 359.714 0
    n(0.002 1,-0.026 2,0.999 3)(0.007 4,-0.024 5,0.999 3)
    μ1.333
    R(-0.009 4,0.001 1,0.014 8)
    T(-349.974 0,-4.178 4,-10.381 8)
    Table 2. Left and right camera calibration parameters
    NumberExperimental value/mmError/mmRMS/mm
    1262.83.52.6
    2261.11.8
    3262.73.4
    4261.92.6
    5260.51.2
    6261.11.8
    7261.42.1
    8262.22.9
    9262.43.1
    10261.62.3
    Table 3. The measurement data
    Sufeng ZHUANG, Yong JI, Dawei TU, Xu ZHANG. Underwater RGB-D Camera Based on Binocular Stereo Vision[J]. Acta Photonica Sinica, 2022, 51(4): 0404003
    Download Citation