Abstract
This paper presents a new approach to accelerate large-scale bundle adjustment (BA). As starting point, we empirically demonstrate that there exist a significant amount of redundant information that has yet to be leveraged in many real-world BA datasets. We propose an adaptive algorithm that builds on this redundancy and further utilizes the bipartite dependency structure of BA. Our algorithm maintains a small set of “training” 3D landmarks that is used to update the camera parameters, while the remaining landmarks are updated using faster point iterations (PI) in an inexact manner. This training set of landmarks is extended as necessary as indicated by a suitable “overfitting” criterion. The proposed algorithm also works gracefully for BA instances with robustified costs (such as the robust Cauchy and Geman-McClure costs). Experimental results on several large-scale BA datasets show that our method achieves faster convergence rates compared to several existing standard approaches.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
as verified in the supplementary material.
- 2.
In principle, the theory such as [35] suggests the necessary accuracy for solving the linear system, but the underlying theory only holds for instances with zero optimal objective value.
- 3.
We observe similar behavior when using only the block-diagonal part in the augmented normal equation.
- 4.
- 5.
- 6.
- 7.
C++ code made available by the authors.
References
Agarwal, S., Snavely, N., Seitz, S.M., Szeliski, R.: Bundle adjustment in the large. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010. LNCS, vol. 6312, pp. 29–42. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-15552-9_3
Bergou, E., Gratton, S., Vicente, L.N.: Levenberg-marquardt methods based on probabilistic gradient models and inexact subproblem solution, with application to data assimilation. SIAM/ASA J. Uncertainty Quantification 4(1), 924–951 (2016)
Bollapragada, R., Nocedal, J., Mudigere, D., Shi, H.J., Tang, P.T.P.: A progressive batching L-BFGS method for machine learning. In: International Conference on Machine Learning, pp. 620–629 (2018)
Byrd, R.H., Hansen, S.L., Nocedal, J., Singer, Y.: A stochastic quasi-newton method for large-scale optimization. SIAM J. Optim. 26(2), 1008–1031 (2016)
Byröd, M., Åström, K.: Bundle adjustment using conjugate gradients with multiscale preconditioning
Byröd, M., Åström, K.: Conjugate gradient bundle adjustment. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010. LNCS, vol. 6312, pp. 114–127. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-15552-9_9
Curtis, F.E., Shi, R.: A fully stochastic second-order trust region method. arXiv preprint arXiv:1911.06920 (2019)
Dai, A., Nießner, M., Zollhöfer, M., Izadi, S., Theobalt, C.: BundleFusion: real-time globally consistent 3D reconstruction using on-the-fly surface reintegration. ACM Trans. Graph. (ToG) 36(4), 1 (2017)
Dan, H., Yamashita, N., Fukushima, M.: Convergence properties of the inexact levenberg-marquardt method under local error bound conditions. Optim. Methods Softw. 17(4), 605–626 (2002)
Davis, T.A., Gilbert, J.R., Larimore, S.I., Ng, E.G.: Algorithm 836: colamd, a column approximate minimum degree ordering algorithm. ACM Trans. Math. Softw. (TOMS) 30(3), 377–380 (2004)
Dellaert, F., Carlson, J., Ila, V., Ni, K., Thorpe, C.E.: Subgraph-preconditioned conjugate gradients for large scale slam. In: 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 2566–2571. IEEE (2010)
Dolan, E.D., Moré, J.J.: Benchmarking optimization software with performance profiles. Math. Program. 91(2), 201–213 (2002)
Enqvist, O., Kahl, F., Olsson, C.: Non-sequential structure from motion. In: 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops), pp. 264–271. IEEE (2011)
Eriksson, A., Bastian, J., Chin, T.J., Isaksson, M.: A consensus-based framework for distributed bundle adjustment. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1754–1762 (2016)
Facchinei, F., Kanzow, C.: A nonsmooth inexact newton method for the solution of large-scale nonlinear complementarity problems. Math. Program. 76(3), 493–512 (1997)
Golub, G.H., Pereyra, V.: The differentiation of pseudo-inverses and nonlinear least squares problems whose variables separate. SIAM J. Numer. Anal. 10(2), 413–432 (1973)
Jeong, Y., Nister, D., Steedly, D., Szeliski, R., Kweon, I.S.: Pushing the envelope of modern methods for bundle adjustment. IEEE Trans. Pattern Anal. Mach. Intell. 34(8), 1605–1617 (2011)
Katyan, S., Das, S., Kumar, P.: Two-grid preconditioned solver for bundle adjustment. In: The IEEE Winter Conference on Applications of Computer Vision, pp. 3599–3606 (2020)
Konolige, T., Brown, J.: Multigrid for bundle adjustment. arXiv preprint arXiv:2007.01941 (2020)
Kushal, A., Agarwal, S.: Visibility based preconditioning for bundle adjustment. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition, pp. 1442–1449. IEEE (2012)
Le, H., Zach, C., Rosten, E., Woodford, O.J.: Progressive batching for efficient non-linear least squares (2020)
Liu, D.C., Nocedal, J.: On the limited memory BFGS method for large scale optimization. Math. Program. 45(1–3), 503–528 (1989)
Lourakis, M., Argyros, A.A.: Is levenberg-marquardt the most efficient optimization algorithm for implementing bundle adjustment? In: Tenth IEEE International Conference on Computer Vision (ICCV 2005) Volume 1. vol. 2, pp. 1526–1531. IEEE (2005)
Mitra, K., Chellappa, R.: A scalable projective bundle adjustment algorithm using the l infinity norm. In: 2008 Sixth Indian Conference on Computer Vision, Graphics & Image Processing, pp. 79–86. IEEE (2008)
Moré, J.J.: The Levenberg-marquardt algorithm: implementation and theory. In: Watson, G.A. (ed.) Numerical Analysis. LNM, vol. 630, pp. 105–116. Springer, Heidelberg (1978). https://doi.org/10.1007/BFb0067700
Mur-Artal, R., Montiel, J.M.M., Tardos, J.D.: ORB-SLAM: a versatile and accurate monocular slam system. IEEE Trans. Robot. 31(5), 1147–1163 (2015)
Ni, K., Steedly, D., Dellaert, F.: Out-of-core bundle adjustment for large-scale 3D reconstruction. In: 2007 IEEE 11th International Conference on Computer Vision, pp. 1–8. IEEE (2007)
Nocedal, J., Wright, S.: Numerical optimization. Springer Science & Business Media (2006)
Parra, Á., Chin, T.J., Eriksson, A., Reid, I.: Visual slam: Why bundle adjust? arXiv preprint arXiv:1902.03747 (2019)
Saad, Y.: Iterative methods for sparse linear systems. SIAM (2003)
Schonberger, J.L., Frahm, J.M.: Structure-from-motion revisited. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4104–4113 (2016)
Tran-Dinh, Q., Pham, N.H., Nguyen, L.M.: Stochastic gauss-newton algorithms for nonconvex compositional optimization. arXiv preprint arXiv:2002.07290 (2020)
Transtrum, M.K., Machta, B.B., Sethna, J.P.: Geometry of nonlinear least squares with applications to sloppy models and optimization. Phys. Rev. E 83(3), 036701 (2011)
Triggs, B., McLauchlan, P.F., Hartley, R.I., Fitzgibbon, A.W.: Bundle adjustment — a modern synthesis. In: Triggs, B., Zisserman, A., Szeliski, R. (eds.) IWVA 1999. LNCS, vol. 1883, pp. 298–372. Springer, Heidelberg (2000). https://doi.org/10.1007/3-540-44480-7_21
Wright, S., Holt, J.N.: An inexact levenberg-marquardt method for large sparse nonlinear least squares. ANZIAM J. 26(4), 387–403 (1985)
Wu, C., Agarwal, S., Curless, B., Seitz, S.M.: Multicore bundle adjustment. In: CVPR 2011, pp. 3057–3064. IEEE (2011)
Zhang, F.: The Schur Complement and its Applications. vol. 4. Springer Science & Business Media (2006). https://doi.org/10.1007/b105056
Zhang, R., Zhu, S., Fang, T., Quan, L.: Distributed very large scale bundle adjustment by global camera consensus. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 29–38 (2017)
Zhou, L., et al.: Stochastic bundle adjustment for efficient and scalable 3D reconstruction. arXiv preprint arXiv:2008.00446 (2020)
Acknowledgement
This work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Zach, C., Le, H. (2023). Exploiting Redundancy for Large Scale Bundle Adjustment: In Partial Defense of Minimization by Alternation. In: Gade, R., Felsberg, M., Kämäräinen, JK. (eds) Image Analysis. SCIA 2023. Lecture Notes in Computer Science, vol 13886. Springer, Cham. https://doi.org/10.1007/978-3-031-31438-4_33
Download citation
DOI: https://doi.org/10.1007/978-3-031-31438-4_33
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-31437-7
Online ISBN: 978-3-031-31438-4
eBook Packages: Computer ScienceComputer Science (R0)