Laser & Optoelectronics Progress, Volume. 56, Issue 15, 151501(2019)
Non-Local Stereo Matching Algorithm Based on Edge Constraint Iteration
Figures & Tables(9)
Fig. 2. First cost aggregation based on minimum spanning tree. (a) Cost aggregation from bottom to up; (b) cost aggregation from up to bottom
Fig. 3. Constraint-based second cost aggregation. (a) Cost aggregation from bottom to up; (b) cost aggregation from up to bottom
Fig. 4. Comparison of disparity maps of two aggregation algorithms. (a) Reference images; (b) disparity maps obtained by original aggregation algorithm; (c) disparity maps obtained by proposed aggregation algorithm
Fig. 5. Experimental results corresponding to parameter Π setting. (a) Mismatching rate of images of group 1 under different parameters; (b) mismatching rate of images of group 2 under different parameters
Fig. 6. Experimental results of Middlebury test dataset (rich texture region). (a) Left of images to be tested; (b) real disparity maps; (c) disparity maps obtained by proposed algorithm
Fig. 7. Experimental results of Middlebury test dataset (low-texture region). (a)Reference images; (b) real disparity maps; (c) disparity maps obtained by proposed algorithm
Fig. 8. Disparity maps obtained by six algorithms. (a) MST; (b) ST-2; (c) CSMST; (d) WCPSP; (e) MST-CD2; (f) proposed algorithm
Table 1. Mismatching rates of six methods in 31 groups of images
View tableTable 1. Mismatching rates of six methods in 31 groups of images
Stereopairs R Mismatching rate /% MST ST-2 CS-MST WCPSP MST-CD2 Proposed algorithm Tsukuba 0.0526 1.493 1.534 1.775 3.016 1.341 1.432 Venus 0.0521 0.253 0.364 1.246 0.865 0.212 0.171 Teddy 0.038 6.014 6.586 5.732 3.741 5.923 6.185 Cones 0.026 2.872 2.883 4.426 3.995 3.224 2.801 Aloe 0.0193 5.025 4.424 4.886 4.082 4.193 3.771 Art 0.0346 10.245 9.984 10.696 8.382 9.043 7.841 Baby1 0.0131 10.506 4.202 8.214 3.341 8.495 4.963 Baby2 0.0332 17.966 15.965 13.543 3.251 15.284 10.542 Baby3 0.0405 7.346 5.162 5.593 2.601 5.984 6.405 Books 0.0336 10.095 10.034 10.666 6.431 9.673 8.882 Bowling1 0.0274 22.566 21.725 19.563 11.371 18.492 21.354 Bowling2 0.0336 11.606 11.045 10.114 5.861 9.623 6.302 Cloth1 0.0103 0.544 0.513 0.635 0.696 0.452 0.341 Cloth2 0.0262 4.195 3.594 4.356 2.441 3.013 2.512 Cloth3 0.0386 2.165 1.631 2.906 1.642 1.663 1.714 Cloth4 0.0213 1.504 1.293 1.886 1.675 1.102 0.871 Dolls 0.0461 6.166 5.104 5.895 4.071 4.922 5.063 Flowerpots 0.0325 22.266 15.423 16.794 13.551 18.525 14.902 Lampshade1 0.0081 12.816 11.715 10.102 10.333 10.444 8.861 Lampshade2 0.0077 12.295 13.306 12.084 7.261 10.753 9.202 Laundry 0.0308 11.423 11.935 11.924 12.076 10.941 11.222 Midd1 0.0429 23.155 21.232 24.434 28.836 22.493 20.361 Midd2 0.0399 32.766 20.411 20.572 35.145 30.874 24.263 Moebius 0.0259 7.875 7.644 7.573 10.536 7.381 7.532 Monopoly 0.0149 20.254 19.031 21.035 28.166 19.453 19.122 Plastic 0.0336 46.696 38.772 45.024 43.253 45.885 34.781 Reindeer 0.044 9.675 7.122 9.796 6.121 8.804 7.373 Rocks1 0.0281 2.765 2.354 3.356 1.792 2.253 1.621 Rocks2 0.0284 2.035 1.664 2.286 1.351 1.603 1.422 Wood1 0.0101 11.926 4.872 10.184 2.851 10.305 6.523 Wood2 0.0188 1.104 2.825 3.176 0.782 0.893 0.731 Average error /% — 10.896 9.173 10.015 8.692 9.784 8.351 Average rank — 4.716 3.684 4.685 2.972 3.263 2.121 Average time /s — 0.781 1.172 4.136 3.545 2.164 1.743
Tools
Get Citation
Copy Citation Text
Ying Luo, Guanying Huo, Jinxin Xu, Qingwu Li. Non-Local Stereo Matching Algorithm Based on Edge Constraint Iteration[J]. Laser & Optoelectronics Progress, 2019, 56(15): 151501
Paper Information
Category: Machine Vision
Received: Jan. 18, 2019
Accepted: Feb. 27, 2019
Published Online: Aug. 5, 2019
The Author Email: Guanying Huo (huoguanying@163.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20