• Acta Optica Sinica
  • Vol. 37, Issue 12, 1215007 (2017)
Jinxin Xu1, Qingwu Li1、2、*, Yan Liu1、2, and Yifei You1
Author Affiliations
  • 1 College of Internet of Things Engineering, Hohai University, Changzhou, Jiangsu 213022, China
  • 2 Changzhou Key Laboratory of Sensor Networks and Environmental Sensing, Changzhou, Jiangsu 213002, China
  • show less
    DOI: 10.3788/AOS201737.1215007 Cite this Article Set citation alerts
    Jinxin Xu, Qingwu Li, Yan Liu, Yifei You. Stereo Matching Algorithm Based on Color Weights and Tree Dynamic Programming[J]. Acta Optica Sinica, 2017, 37(12): 1215007 Copy Citation Text show less
    Disparity maps computed by AD-Census and proposed algorithms. (a) Reference image; (b) real disparity map; (c) AD-Census; (d) proposed algorithm
    Fig. 1. Disparity maps computed by AD-Census and proposed algorithms. (a) Reference image; (b) real disparity map; (c) AD-Census; (d) proposed algorithm
    Cost aggregation strategy based on color weights. (a) Cost aggregation process; (b) row (column) aggregation process
    Fig. 2. Cost aggregation strategy based on color weights. (a) Cost aggregation process; (b) row (column) aggregation process
    (a) Reference images; (b)disparity maps of original cross-based aggregation; (c) disparity maps of proposed algorithm
    Fig. 3. (a) Reference images; (b)disparity maps of original cross-based aggregation; (c) disparity maps of proposed algorithm
    Tree structures. (a) Horizontal tree; (b) vertical tree
    Fig. 4. Tree structures. (a) Horizontal tree; (b) vertical tree
    Disparity maps computed by different algorithms. (a) Reference images; (b) WTA algorithm; (c) DP algorithm; (d) proposed algorithm
    Fig. 5. Disparity maps computed by different algorithms. (a) Reference images; (b) WTA algorithm; (c) DP algorithm; (d) proposed algorithm
    Comparison of mismatching rates among three algorithms in non-occluded regions
    Fig. 6. Comparison of mismatching rates among three algorithms in non-occluded regions
    Experimental results of the Middlebury benchmark images. (a) Reference images; (b) real disparity maps; (c) disparity maps obtained by proposed algorithm
    Fig. 7. Experimental results of the Middlebury benchmark images. (a) Reference images; (b) real disparity maps; (c) disparity maps obtained by proposed algorithm
    Disparity maps obtained by five algorithms. (a) CrossTrees+SP algorithm; (b) CostFilter algorithm; (c) TwoStep algorithm; (d) AdaptAggrDP algorithm; (e) proposed algorithm
    Fig. 8. Disparity maps obtained by five algorithms. (a) CrossTrees+SP algorithm; (b) CostFilter algorithm; (c) TwoStep algorithm; (d) AdaptAggrDP algorithm; (e) proposed algorithm
    AlgorithmTsukubaVenusTeddyConesAverage error (Nonocc)
    NonoccAllDiscNonoccAllDiscNonoccAllDiscNonoccAllDisc
    Proposed1.252.186.720.240.843.275.5410.5014.902.779.027.982.45
    CostFilter1.521.857.610.200.392.426.1611.8016.002.718.247.662.65
    CrossTrees+SP1.681.997.820.220.322.846.2311.7014.802.527.717.502.66
    TwoStep2.913.6813.30.270.452.637.4212.6018.004.0910.1010.303.67
    AdaptAggrDP1.573.508.271.532.6912.406.7914.3016.205.5313.2014.803.86
    Table 1. Mismatching rates of five algorithms%
    Jinxin Xu, Qingwu Li, Yan Liu, Yifei You. Stereo Matching Algorithm Based on Color Weights and Tree Dynamic Programming[J]. Acta Optica Sinica, 2017, 37(12): 1215007
    Download Citation