Chinese Optics, Volume. 17, Issue 2, 278(2024)

Stereo matching algorithm based on multi-feature SAD-Census transformation

Fu-pei WU, Geng-nan HUANG*, Yu-hao LIU, Wei-lin YE, and Sheng-ping LI
Author Affiliations
  • Department of Mechanical Engineering, College of Engineering, Shantou University, Shantou 515063, China
  • show less
    Figures & Tables(16)
    Schematic diagram of improved Census transform
    Adaptive crossing domain
    Schematic diagram of optimized aggregation path
    Parallax maps of three cost calculation methods. (a) Traditional SAD-Census algorithms; (b) traditional SAD + improved Census algorithm; (a) the proposed algorithm
    Experimental results of 4 kinds of algorithms
    Mismatch rate of different algorithms
    Experimental results of 4 kinds of algorithms at different brightness levels
    Experimental results of different algorithms under noise interference
    Comparison results of distance measuring experiments
    Edge extraction of sample parallax result maps
    Gauge block parallax map
    • Table 1. Experimental parameter setting

      View table
      View in Article

      Table 1. Experimental parameter setting

      ParameterValueParameterValue
      б2λ0.003
      τ116ε0.001
      τ24ω9×7
      L130β10.6
      L215β20.4
      δ01
    • Table 2. Mismatch rate of different algorithms in the non-occluded region and all regions (%)

      View table
      View in Article

      Table 2. Mismatch rate of different algorithms in the non-occluded region and all regions (%)

      AlgorithmTsukubaVenusTeddyConesAvg
      NonallDiscNonallDiscNonallDiscNonallDiscNonallDisc
      GC-occ1.192.016.241.642.196.7511.2017.4019.805.3612.4013.004.848.5011.45
      SemiGlob3.263.9612.801.001.5711.306.0212.2016.303.069.758.903.346.8712.32
      RTCensus5.086.2519.201.582.4214.207.9613.820.304.109.5412.204.688.0116.48
      HCA1.312.470.940.110.230.244.239.453.675.6411.594.302.835.942.29
      Cross-ScaleGF2.382.858.401.131.989.267.0514.916.803.3211.007.993.477.6810.61
      GradAdaptWgt2.262.638.990.991.394.928.0013.1018.602.617.677.433.466.199.99
      Proposed algorithm1.252.730.900.150.310.314.628.732.984.6910.984.112.675.692.08
    • Table 3. Experimental results of three-dimensional distance measurement

      View table
      View in Article

      Table 3. Experimental results of three-dimensional distance measurement

      序号实际距离/mm测量距离/mm距离误差/mm距离误差率/%
      1200198.42−1.580.79
      2300302.552.550.85
      3400403.763.760.94
      4500495.05−4.950.99
      5600606.066.061.01
      6700707.567.561.08
      7800809.529.521.19
      8900912.2412.241.36
    • Table 4. Sample experimental results

      View table
      View in Article

      Table 4. Sample experimental results

      样本序号实际长度/mm测量长度/mm误差/mm误差率/%
      12019.9575−0.0425−0.21
      3534.6144−0.3856−1.1
      23535.16490.1649−0.47
      5049.9672−0.0328−0.06
      354.9439−0.0561−1.12
      1514.7565−0.2435−1.62
      42020.35410.35410.71
      3029.5784−0.4216−1.41
    • Table 5. Thickness measurement results

      View table
      View in Article

      Table 5. Thickness measurement results

      序号实际厚度/mm测量厚度/mm测量误差/mm绝对误差率/%
      11.001.01050.01051.05
      21.101.12350.02352.13
      31.201.23350.03352.79
      41.501.55630.05633.75
      52.001.9346−0.06543.27
    Tools

    Get Citation

    Copy Citation Text

    Fu-pei WU, Geng-nan HUANG, Yu-hao LIU, Wei-lin YE, Sheng-ping LI. Stereo matching algorithm based on multi-feature SAD-Census transformation[J]. Chinese Optics, 2024, 17(2): 278

    Download Citation

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

    Category:

    Received: May. 8, 2023

    Accepted: --

    Published Online: Apr. 15, 2024

    The Author Email:

    DOI:10.37188/CO.2023-0082

    Topics