• Laser & Optoelectronics Progress
  • Vol. 58, Issue 12, 1215003 (2021)
Jianhong Zhu, Caosong Wang*, and Meifeng Gao
Author Affiliations
  • Key Laboratory of Advanced Process Control for Light Industry, Ministry of Education, Jiangnan University, Wuxi, Jiangsu 214122, China
  • show less
    DOI: 10.3788/LOP202158.1215003 Cite this Article Set citation alerts
    Jianhong Zhu, Caosong Wang, Meifeng Gao. An Improved Matching Algorithm of Census Transform and Adaptive Window[J]. Laser & Optoelectronics Progress, 2021, 58(12): 1215003 Copy Citation Text show less
    Traditional Census transform and improved Census transform. (a) Original image of Teddy; (b) traditional Census transform result; (c) improved Census transform result
    Fig. 1. Traditional Census transform and improved Census transform. (a) Original image of Teddy; (b) traditional Census transform result; (c) improved Census transform result
    Comparison before and after aggregation. (a) Before aggregation; (b) after aggregation
    Fig. 2. Comparison before and after aggregation. (a) Before aggregation; (b) after aggregation
    Experimental results. (a) Original left images; (b) true disparity images; (c) traditional Census transform results; (d) proposed algorithm results
    Fig. 3. Experimental results. (a) Original left images; (b) true disparity images; (c) traditional Census transform results; (d) proposed algorithm results
    Real scene results. (a) Actual left image; (b) actual right image; (c) disparity map
    Fig. 4. Real scene results. (a) Actual left image; (b) actual right image; (c) disparity map
    ParameterT1T2T3
    Value50010005000
    Table 1. Adaptive variance threshold parameter
    SampleProposed algorithm /sTraditional algorithm /sReduction /%
    Venus3.424.4730.70
    Teddy3.915.1331.20
    Cones2.874.1745.30
    Tsukba3.244.5139.19
    Table 2. Time-consuming comparison of proposed algorithm and traditional algorithm
    AlgorithmVenusTeddyConesTsukbaAverage
    NoccAllNoccAllNoccAllNoccAll
    Ada_SGM4.094.2315.6819.8814.1318.3110.3010.5412.14
    RINCensus1.111.769.7617.308.0916.204.786.008.13
    ADCensus0.090.254.010.252.427.251.071.482.10
    NPCT0.840.956.3511.725.8710.334.545.985.82
    SemiGlob1.001.576.0212.203.069.753.263.965.10
    Proposed1.092.316.7713.013.068.384.355.135.51
    Table 3. Error matching rate of different algorithms unit:%
    Actual distance/mmMeasuring distance/mmError/%
    500506.051.21
    600610.801.80
    700714.762.10
    800821.842.73
    900929.793.31
    10001037.703.77
    Table 4. Actual distance and measuring distance
    Jianhong Zhu, Caosong Wang, Meifeng Gao. An Improved Matching Algorithm of Census Transform and Adaptive Window[J]. Laser & Optoelectronics Progress, 2021, 58(12): 1215003
    Download Citation