Acta Photonica Sinica, Volume. 53, Issue 8, 0810003(2024)
Dynamic Weight Cost Aggregation Algorithm for Stereo Matching Based on Adaptive Window
Figures & Tables(21)
Fig. 6. Influence of α and β on error matching rate of disparity images
Fig. 9. Experimental results of proposed algorithm on Middlebury image pairs
Fig. 13. The reconstruction results of four surface types using different algorithms
Table 1. Experimental parameter settings
View tableView in ArticleTable 1. Experimental parameter settings
Parameter Value Parameter Value λAD 10 λCensus 30 τ1 15 a 100 τ2 26 δ 20 L1 50 α 4 L2 25 β 11 τS 20 τH 0.4
Table 2. Average error matching rate in non-occlusion region and all region（%）
View tableView in ArticleTable 2. Average error matching rate in non-occlusion region and all region（%）
Algorithm Teddy Baby1 Cloth3 Wood2 Average n-occ all n-occ all n-occ all n-occ all Mei's aggregation 5.91 8.84 6.54 7.58 3.18 3.96 6.88 7.93 6.35 Proposed algorithm 4.93 7.08 4.78 6.88 2.63 3.01 1.83 2.54 4.21
Table 3. Error matching rates of different algorithms in non-occlusion region and all region （%）
View tableView in ArticleTable 3. Error matching rates of different algorithms in non-occlusion region and all region （%）
Algorithm Teddy Baby1 Baby2 Baby3 Flowerpots Average n-occ all n-occ all n-occ all n-occ all n-occ all n-occ all LSECVR 4.91 16.08 5.12 8.80 6.61 9.79 7.61 14.48 14.66 19.83 7.78 13.80 GF 5.77 9.62 6.53 9.78 7.97 10.53 5.70 8.58 10.64 14.41 7.32 10.58 Semiglob 6.25 12.36 7.36 10.98 9.41 14.98 7.13 10.33 12.77 17.34 8.58 13.20 AD-Census 4.93 7.08 4.78 6.88 7.33 8.13 6.85 8.97 9.12 12.88 6.60 8.79 Patchmatch 3.56 5.03 5.38 6.53 6.23 7.86 6.44 7.01 8.56 10.68 6.03 7.42 Proposed algorithm 2.52 3.16 3.86 4.98 4.12 6.08 3.43 5.26 6.63 8.79 4.12 5.65
Table 4. Execute time of six algorithms （s）
View tableView in ArticleTable 4. Execute time of six algorithms （s）
Algorithm LSECVR GF Semiglob AD-Census Patchmatch Proposed Execute time/s 1.57 1.96 0.98 1.56 30.84 2.53
Table 5. Camera parameters
View tableView in ArticleTable 5. Camera parameters
Parameter matrix Left camera Right camera Inner parameter matrix Distortion coefficient Rotation matrix Translation vector
Table 6. Experimental results of three-dimensional reconstruction
View tableView in ArticleTable 6. Experimental results of three-dimensional reconstruction
Convex surface Step surface Angular surface Concave surface Sampleimage Acquireimage Reconstructionimage
Table 7. Error measuring rates of sample in length
View tableView in ArticleTable 7. Error measuring rates of sample in length
Name Actual length/mm Measure length/mm Error/mm Error rate/% Convex surface 16 15.873 2 -0.126 8 -0.793 8 7.974 4 -0.025 6 -0.320 Step surface 25 25.299 8 0.299 8 1.199 10 9.954 3 -0.045 7 -0.457 Angular surface 16 16.157 5 0.157 5 0.984 8 8.047 7 0.047 7 0.596 Concave surface 20 19.836 9 -0.163 1 -0.815 10 9.962 4 -0.037 6 -0.376
Table 8. Error measuring rates of sample in height
View tableView in ArticleTable 8. Error measuring rates of sample in height
Name Maximum error/mm Average error/mm Average absolute error/mm SD error/mm Convex surface 0.233 4 0.081 2 0.112 3 0.098 1 Step surface 0.144 8 -0.100 4 0.061 3 0.066 1 Angular surface 0.221 7 -0.007 4 0.036 8 0.051 8 Concave surface 0.336 3 -0.054 1 0.103 5 0.091 7 Ave 0.234 1 -0.020 2 0.078 5 0.076 9
Tools
Get Citation
Copy Citation Text
Fupei WU, Yuhao LIU, Rui WANG, Shengping LI. Dynamic Weight Cost Aggregation Algorithm for Stereo Matching Based on Adaptive Window[J]. Acta Photonica Sinica, 2024, 53(8): 0810003
Paper Information
Category:
Received: Jan. 24, 2024
Accepted: Mar. 21, 2024
Published Online: Oct. 15, 2024
The Author Email: LI Shengping (spli@stu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20