To address the low accuracy of traditional stereo matching methods in depth-discontinuous and weak-texture regions, we propose an improved algorithm based on dynamic multi-feature fusion and dual-branch adaptive aggregation. A nonlinear weighting function dynamically integrates noise-resistant Re-Census, multi-directional gradient and Lab-color costs. Distinct arm extension rules are applied to weak texture and depth-discontinuous areas, enabling a dual-branch adaptive aggregation that adapts to local scene characteristics. Disparity estimation follows a winner takes all (WTA) strategy. The ultimate disparity map is generated through a region-based disparity optimization module and multiple optimization processes. Experimental results on the Middlebury dataset indicate a 16.1% relative decrease in the mismatch rate compared to the baseline. Evaluations on the KITTI dataset and real-world scenes captured with a ZED 2i camera further demonstrate the robustness of the proposed algorithm.
| [1] |
Tosi F, Bartolomei L, Poggi M. A survey on deep stereo matching in the twenties. International journal of computer vision, 2025, 133(7): 4245-4276 J]
|
| [2] |
Jia X, Yang S, Guan D. Semantics-aware transformer for 3D reconstruction from binocular images. Optoelectronics letters, 2022, 18(5): 293-299 J]
|
| [3] |
Zhao Y, Qin H, Xu L, et al.. A review of deep learning-based stereo vision techniques for phenotype feature and behavioral analysis of fish in aquaculture. Artificial intelligence review, 2025, 58(1): 1-61[J]
|
| [4] |
Meng L, Hu Y, Guo J, et al.. Intelligent design of biotechnology ceramic structures using 3D structured light simulation technology for commercial applications. Journal of commercial biotechnology, 2025, 30(2): 234-247[J]
|
| [5] |
Luo L, Liu S, Li F, et al.. Three-dimensional vibration visual measurement method based on binocular camera. Archives of advanced engineering science, 2026, 4(1): 78-88 J]
|
| [6] |
Li D, Yu J, Du Z, et al.. Advances in the application of stereo vision in aquaculture with emphasis on fish: a review. Reviews in aquaculture, 2024, 16(4): 1718-1740 J]
|
| [7] |
Lei J F, Wei Z H, Li Y L, et al.. Adaptive stereo matching based on local information entropy and improved AD-Census transform. Laser & optoelectronics progress, 2025, 62(2): 205-213[J]
|
| [8] |
Yang C Y, Song Z R, Zhang X. A stereo matching algorithm for coal mine underground images based on threshold and weight under census transform. Coal science and technology, 2024, 52(6): 216-225[J]
|
| [9] |
Li T L, Huang Z, He L L, et al.. Stereo matching algorithm based on improved census transform and adaptive window optimization. 2024 4th International Conference on Computer Science and Blockchain, September 6–8, 2024, Shenzhen, China, 2024, Shenzhen, Shenzhen University, 219-224[C]
|
| [10] |
Li H, Huang M H. Adaptive region division stereo matching algorithm. Laser & optoelectronics progress, 2023, 60(10): 172-182[J]
|
| [11] |
Wu W, Xu X, Wang W, et al.. Stereo matching method with cost volume collaborative filtering. Electronics, 2024, 13(15): 3095 J]
|
| [12] |
Chang J R, Chen Y S. Pyramid stereo matching network. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18–23, 2018, Salt Lake City, UT, USA, 2018, New York, IEEE, 5410-5418[C]
|
| [13] |
Dai Z, Geng Y, Xue M, et al.. ZHNet: improved deep stereo matching network based on PSMNet. 2024 2nd International Conference on Intelligent Perception and Computer Vision, May 17–19, 2024, Xiamen, China, 2024, New York, IEEE, 57-61[C]
|
| [14] |
Xu G, Wang Y, Cheng J, et al.. Accurate and efficient stereo matching via attention concatenation volume. IEEE transactions on pattern analysis and machine intelligence, 2024, 46(4): 2461-2474 J]
|
| [15] |
CHENG J, LIU L, XU G, et al. MonSter: marry mono-depth to stereo unleashes power[EB/OL]. (2025-01-15) [2025-05-24]. https://arxiv.org/abs/2501.08643.
|
| [16] |
Wang C, Ding Y, Huang C L, et al.. Bitsliced optimization of SM4 algorithm with the SIMD instruction set. Journal of computer research and development, 2024, 61(8): 2097-2109[J]
|
| [17] |
Lu Z, Wang J, Li Z, et al.. A resource-efficient pipelined architecture for real-time semi-global stereo matching. IEEE transactions on circuits and systems for video technology, 2022, 32(2): 660-673 J]
|
| [18] |
Iqbal U, Davies T, Perez P. A review of recent hardware and software advances in GPU-accelerated edge-computing single-board computers (SBCs) for computer vision. Sensors, 2024, 24(15): 4830 J]
|
| [19] |
Azali M N Z, Hamzah R A, Noh Z M. Disparity map algorithm using census transform and hierarchical segment-tree from stereo image. Engineering Technology International Conference, September 7–8, 2022, Pekan, Malaysia, 2022244-249[C]
|
| [20] |
Wang Y, Zheng J, Zhang C, et al.. DualNet: robust self-supervised stereo matching with pseudo-label supervision. Proceedings of the AAAI Conference on Artificial Intelligence, February 25–March 4, 2025, Philadelphia, PA, USA, 2025, New York, AAAI Press, 8178-8186[C]
|
| [21] |
Zhao H, Zhou H, Zhang Y, et al.. High-frequency stereo matching network. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18–22, 2023, Vancouver, BC, Canada, 2023, New York, IEEE, 1327-1336[C]
|
| [22] |
Chen Y, Yu J Y, Yu H S. Multi-scale fusion stereo matching algorithm based on adaptive texture region. Computer engineering and applications, 2023, 59(18): 198-206[J]
|
| [23] |
Ferrera M, Boulch A, Moras J. Fast stereo disparity maps refinement by fusion of data-based and model-based estimations. Proceedings of the 2019 International Conference on 3D Vision, September 16–19, 2019, Québec City, QC, Canada, 2019, New York, IEEE, 9-17[C]
|
| [24] |
Yang R, Li X, Cong R, et al.. Unsupervised hierarchical iterative tile refinement network with 3D planar segmentation loss. IEEE robotics and automation letters, 2024, 9(3): 2678-2685 J]
|
Rights & permissions
Tianjin University of Technology
PDF
