• Chinese Journal of Lasers
  • Vol. 47, Issue 12, 1204007 (2020)
Fang Wei1、* and Yang Kui2
Author Affiliations
  • 1School of Automation, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • 2School of Instrumentation Science and Opto-Electronics Engineering, Beihang University, Beijing 100191, China
  • show less
    DOI: 10.3788/CJL202047.1204007 Cite this Article Set citation alerts
    Fang Wei, Yang Kui. Inverse Depth Adaptive Weighting Based Multi-View Triangulation Method[J]. Chinese Journal of Lasers, 2020, 47(12): 1204007 Copy Citation Text show less
    Schematic diagram of triangulation description
    Fig. 1. Schematic diagram of triangulation description
    Implementation process of our method
    Fig. 2. Implementation process of our method
    Synthetic datasets for multi-view triangulation. (a) Type A; (b) type B; (c) type C
    Fig. 3. Synthetic datasets for multi-view triangulation. (a) Type A; (b) type B; (c) type C
    Iteration performance at different noise levels
    Fig. 4. Iteration performance at different noise levels
    Multi-view triangulation results of our method under the public datasets. (a) Lund Cathedral; (b) Aos Hus; (c) San Marco; (d) Orebro Castle; (e) Buddah Statue; (f) East Indiaman Goteborg; (g) Ystad Monestary; (h) Round Church; (i) Skansen Kronan; (j) Skansen Lejonet
    Fig. 5. Multi-view triangulation results of our method under the public datasets. (a) Lund Cathedral; (b) Aos Hus; (c) San Marco; (d) Orebro Castle; (e) Buddah Statue; (f) East Indiaman Goteborg; (g) Ystad Monestary; (h) Round Church; (i) Skansen Kronan; (j) Skansen Lejonet
    Number of iterations of our method for multi-view triangulation in public datasets
    Fig. 6. Number of iterations of our method for multi-view triangulation in public datasets
    σ /pixelMethodType AType BType C
    Time /s3D error /m2D error /pixelTime /s3D error /m2D error /pixelTime /s3D error /m2D error /pixel
    Midpoint0.6480.0236.1072.5210.0215.9450.8530.0175.825
    6Ours1.7610.0175.7285.9710.0185.9382.2510.0165.801
    L2Rep2.9960.0155.70411.6680.0175.9364.0050.0165.797
    Midpoint0.6170.04512.2442.4620.05911.9410.8100.03611.652
    12Ours1.8020.03111.6695.6750.03711.8832.0680.03111.595
    L2Rep2.8410.03011.62011.6780.03511.8793.9330.02911.582
    Midpoint0.6030.09425.9532.5780.20824.2780.8490.07823.498
    24Ours1.9140.06323.0097.1910.07423.7812.3820.06423.216
    L2Rep3.2240.05922.90514.9650.07123.7734.0170.05923.180
    Table 1. Results obtained by the multi-view triangulation method based on simulation data
    DatasetTime /sMean 2D reprojection errors /pixel
    IDViewPointMidpointOursL2RepMeanVariance
    MidpointOursL2RepMidpointOursL2Rep
    1120815905523.08243.01491.9101.0881.0781.0770.2170.2060.205
    280035413422.30345.27393.0630.8160.8050.8050.3020.2890.287
    3149823150733.78666.944134.2050.8070.7990.7980.3310.3160.315
    47615385710.99820.41040.5320.9420.9360.9360.1960.1910.190
    53221563569.44015.23331.8730.6510.6490.6490.2700.2660.265
    6179256552.7235.73611.0401.1271.1221.1210.3860.3770.376
    729013995114.52026.32853.4600.9700.9680.9670.1700.1690.168
    892846436.36611.69922.4450.3870.3850.3850.1210.1190.119
    9131283713.8127.25414.7930.8070.8020.8020.1750.1700.169
    10368744238.37015.87134.3501.0321.0231.0220.2090.2020.200
    Table 2. Results of different methods based on Lund's public dataset
    IDTime /sMean 2D reprojection error /pixel
    MidpointOursL2RepMidpointOursL2Rep
    120.50844.27590.5562.1362.1052.103
    223.82946.67396.9192.9642.9342.930
    334.91363.022139.8732.6632.6272.624
    411.38020.58546.9091.6771.6641.662
    59.69117.31638.9843.0503.0463.044
    62.7245.83811.8092.3562.3502.351
    713.90327.21358.1752.2572.2412.236
    86.28912.39726.1472.8292.8132.811
    93.8596.82715.9001.9171.8711.870
    107.78615.55034.1842.1392.1092.105
    Table 3. Experimental results of different methods on the Lund public dataset (σ=5 pixel)
    Fang Wei, Yang Kui. Inverse Depth Adaptive Weighting Based Multi-View Triangulation Method[J]. Chinese Journal of Lasers, 2020, 47(12): 1204007
    Download Citation