Abstract
Sparse bundle adjustment (SBA) is a key but time- and memory-consuming step in three-dimensional (3D) reconstruction. In this paper, we propose a 3D point-based distributed SBA algorithm (DSBA) to improve the speed and scalability of SBA. The algorithm uses an asynchronously distributed sparse bundle adjustment (A-DSBA) to overlap data communication with equation computation. Compared with the synchronous DSBA mechanism (SDSBA), A-DSBA reduces the running time by 46%. The experimental results on several 3D reconstruction datasets reveal that our distributed algorithm running on eight nodes is up to five times faster than that of the stand-alone parallel SBA. Furthermore, the speedup of the proposed algorithm (running on eight nodes with 48 cores) is up to 41 times that of the serial SBA (running on a single node).
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Agarwal S, Furukawa Y, Snavely N, et al., 2011. Building Rome in a day. Commun ACM, 54(10):105–112. https://doi.org/10.1145/2001269.2001293
Choudhary S, Gupta S, Narayanan P, 2010. Practical time bundle adjustment for 3D reconstruction on the GPU. In: Kutulakos KN (Ed.), Trends and Topics in Computer Vision. Springer–Verlag Berlin Heidelberg. https://doi.org/10.1007/978-3-642-35740-4_33
Eriksson A, Bastian J, Chin T, et al., 2016. A consensusbased framework for distributed bundle adjustment. IEEE Conf on Computer Vision and Pattern Recognition, p.1754–1762. https://doi.org/10.1109/CVPR.2016.194
Frahm J, Fite–Georgel P, Gallup D, et al., 2010. Building Rome on a cloudless day. Proc 11th European Conf on Computer Vision, p.368–381. https://doi.org/10.1007/978-3-642-15561-1_27
Hänsch R, Drude I, Hellwich O, 2016. Modern methods of bundle adjustment on the GPU. ISPRS Ann Photogr Remote Sens Spatial Inform Sci, III–(3):43–50. https://doi.org/10.5194/isprs-annals-III-3-43-2016
Hartley R, Zisserman A, 2000. Multiple View Geometry in Computer Vision. Cambridge University Press, Cambridge, p.1865–1872.
Heinecke A, Vaidyanathan K, Smelyanskiy M, et al., 2013. Design and implementation of the linpack benchmark for single and multi–node systems based on IntelR Xeon Phi coprocessor. Proc 27th Int Symp on Parallel and Distributed Processing, p.126–137. https://doi.org/10.1109/IPDPS.2013.113
Husbands P, Yelick K, 2007. Multi–threading and onesided communication in parallel LU factorization. ACM/IEEE Conf on Supercomputing, Article 31. https://doi.org/10.1145/1362622.1362664
Jo G, Nah J, Lee J, et al., 2015. Accelerating LINPACK with MPI–OpenCL on clusters of multi–GPU nodes. IEEE Trans Parall Distrib Syst, 26(7):1814–1825. https://doi.org/10.1109/TPDS.2014.2321742
Li X, Wu C, Zach C, et al., 2008. Modeling and recognition of landmark image collections using iconic scene graphs. Proc 10th European Conf on Computer Vision, p.427–440. https://doi.org/10.1007/978-3-540-88682-2_33
Liu X, Gao W, Hu Z, 2012. Hybrid parallel bundle adjustment for 3D scene reconstruction with massive points. Comput Sci Technol, 27(6):1269–1280. https://doi.org/10.1007/s11390-012-1303-3
Lourakis M, Argyros A, 2009. SBA: a software package for generic sparse bundle adjustment. ACM Trans Math Soft, 36(1), Article 2. https://doi.org/10.1145/1486525.1486527
Ni K, Steedly D, Dellaert F, 2007. Out–of–core bundle adjustment for large–scale 3D reconstruction. Proc 11th Int Conf on Computer Vision, p.1–8. https://doi.org/10.1109/ICCV.2007.4409085
Ramamurthy K, Lin C, Aravkin A, et al., 2017. Distributed bundle adjustment. IEEE Int Conf on Computer Vision Workshop, p.2146–2154. https://doi.org/10.1109/ICCVW.2017.251
Snavely N, Seitz S, Szeliski R, 2006. Photo tourism: exploring photo collections in 3D. ACM Trans Graph, 25(3):835–846. https://doi.org/10.1145/1141911.1141964
Snavely N, Seitz S, Szeliski R, 2008. Skeletal graphs for efficient structure from motion. IEEE Conf on Computer Vision and Pattern Recognition, p.1–8. https://doi.org/10.1109/CVPR.2008.4587678
Triggs B, McLauchlan P, Hartley R, et al., 1999. Bundle adjustment—modern synthesis. Int Workshop on Vision Algorithms: Theory and Practice, p.298–372. https://doi.org/10.1007/3-540-44480-7_21
Wu C, Agarwal S, Curless B, et al., 2011. Multicore bundle adjustment. IEEE Conf on Computer Vision and Pattern Recognition, p.3057–3064. https://doi.org/10.1109/CVPR.2011.5995552
Zheng E, Wu C, 2015. Structure from motion using structureless resection. IEEE Int Conf on Computer Vision, p.2075–2083. https://doi.org/10.1109/ICCV.2015.240
Author information
Authors and Affiliations
Corresponding author
Additional information
Project supported by the National Natural Science Foundation of China (Nos. U1435219, U1435222, and 61572515), the National Key R & D Program of China (No. 2016YFB0200401), and the Major Research Plan of the National Key R & D Program of China (No. 2016YFC0901600)
Rights and permissions
About this article
Cite this article
Shen, Xl., Dou, Y., Mills, S. et al. Distributed sparse bundle adjustment algorithm based on three-dimensional point partition and asynchronous communication. Frontiers Inf Technol Electronic Eng 19, 889–904 (2018). https://doi.org/10.1631/FITEE.1800173
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/FITEE.1800173
Key words
- Sparse bundle adjustment
- Parallel
- Distributed sparse bundle adjustment
- Three-dimensional reconstruction
- Asynchronous