Abstract
This paper tackles the problem of relative pose estimation between two monocular camera images in textureless scenes. Due to a lack of point matches, point-based approaches such as the 5-point algorithm often fail when used in these scenarios. Therefore we investigate relative pose estimation from line observations. We propose a new algorithm in which the relative pose estimation from lines is extended by a 3D line direction estimation step. Using the estimated line directions, the robustness and computational efficiency of the relative pose calculation is greatly improved. Furthermore, we investigate line matching techniques as the quality of the matches influences directly the outcome of the relative pose estimation. We develop a novel line matching strategy for small baseline matching based on optical flow which outperforms current state-of-the-art descriptor-based line matchers. First, we describe in detail the proposed line matching approach. Second, we introduce our relative pose estimation based on 3D line directions. We evaluate the different algorithms on synthetic and real sequences and demonstrate that in the targeted scenarios we outperform the state-of-the-art in both accuracy and computation time.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
A scene complies with the “Manhattan world” assumption if it has three dominant line directions which are orthogonal and w.l.o.g. can be assumed to coincide with the x-, y- and z-axis of the world coordinate system.
- 2.
We thank the authors for providing us with their implementation.
- 3.
We use the implementation from https://github.com/bverhagen/SMSLD/tree/master/MSLD/MSLD/MSLD.
- 4.
We use the implementation of OpenCV 3.0.0.
- 5.
- 6.
We use the implementation in OpenCV 3.0.0 with default parameters.
- 7.
References
Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Int. J. Comput. Vis. 60, 91–110 (2004)
Fischler, M.A., Bolles, R.C.: Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography. Commun. ACM 24, 381–395 (1981)
Nister, D.: An efficient solution to the five-point relative pose problem. IEEE Trans. Pattern Anal. Mach. Intell. 26, 756–770 (2004)
von Schmude, N., Lothe, P., Jähne, B.: Relative pose estimation from straight lines using parallel line clustering and its application to monocular visual odometry. In: International Conference on Computer Vision Theory and Applications, pp. 421–431 (2016)
Hartley, R.I.: Projective reconstruction from line correspondences. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 903–907 (1994)
Hartley, R.I., Zisserman, A.: Multiple View Geometry in Computer Vision, 2nd edn. Cambridge University Press, Cambridge (2004). ISBN 0521540518
Weng, J., Huang, T., Ahuja, N.: Motion and structure from line correspondences; closed-form solution, uniqueness, and optimization. IEEE Trans. Pattern Anal. Mach. Intell. 14, 318–336 (1992)
Elqursh, A., Elgammal, A.: Line-based relative pose estimation. In: 2011 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3049–3056 (2011)
Wang, G., Wu, J., Ji, Z.: Single view based pose estimation from circle or parallel lines. Pattern Recogn. Lett. 29, 977–985 (2008)
Bazin, J., Demonceaux, C., Vasseur, P., Kweon, I.: Motion estimation by decoupling rotation and translation in catadioptric vision. Comput. Vis. Image Underst. 114, 254–273 (2010)
Wang, Z., Wu, F., Hu, Z.: MSLD: a robust descriptor for line matching. Pattern Recogn. 42, 941–953 (2009)
Hirose, K., Saito, H.: Fast line description for line-based slam. In Bowden, R., Collomosse, J., Mikolajczyk, K. (eds.) 2012 British Machine Vision Conference, pp. 83.1–83.11 (2012)
Zhang, L., Koch, R.: Line matching using appearance similarities and geometric constraints. In: Pinz, A., Pock, T., Bischof, H., Leberl, F. (eds.) DAGM/OAGM 2012. LNCS, vol. 7476, pp. 236–245. Springer, Heidelberg (2012). doi:10.1007/978-3-642-32717-9_24
Zhang, L., Koch, R.: An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency. J. Vis. Commun. Image Represent. 24, 794–805 (2013)
Deriche, R., Faugeras, O.: Tracking line segments. In: Faugeras, O. (ed.) ECCV 1990. LNCS, vol. 427, pp. 259–268. Springer, Heidelberg (1990). doi:10.1007/BFb0014872
Chiba, N., Kanade, T.: A tracker for broken and closely-spaced lines. In. ISPRS International Society for Photogrammetry and Remote Sensing Conference, Hakodate, Japan, pp. 676–683 (1998)
Bartoli, A., Sturm, P.: Structure-from-motion using lines: representation, triangulation, and bundle adjustment. Comput. Vis. Image Underst. 100, 416–441 (2005)
Schindler, G., Krishnamurthy, P., Dellaert, F.: Line-based structure from motion for urban environments. In: Third International Symposium on 3D Data Processing, Visualization, and Transmission (3DPVT 2006), pp. 846–853 (2006)
Zhou, H., Zou, D., Pei, L., Ying, R., Liu, P., Yu, W.: Structslam: visual slam with building structure lines. IEEE Trans. Veh. Technol. 1 (2015)
Witt, J., Weltin, U.: Robust stereo visual odometry using iterative closest multiple lines. In: 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2013), pp. 4164–4171 (2013)
Holzmann, T., Fraundorfer, F., Bischof, H.: Direct stereo visual odometry based on lines. In: International Conference on Computer Vision Theory and Applications, pp. 474–485 (2016)
Lucas, B.D., Kanade, T.: An iterative image registration technique with an application to stereo vision. IJCAI 81, 674–679 (1981)
Shi, J., Tomasi, C.: Good features to track. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 593–600 (1994)
Gower, J.C., Dijksterhuis, G.B.: Procrustes Problems, vol. 3. Oxford University Press, Oxford (2004)
Umeyama, S.: Least-squares estimation of transformation parameters between two point patterns. IEEE Trans. Pattern Anal. Mach. Intell. 13, 376–380 (1991)
Dempster, A.P., Laird, N.M., Rubin, D.B.: Maximum likelihood from incomplete data via the EM algorithm. J. Roy. Stat. Soc. Ser. B (Methodol.) 1–38 (1977)
Antone, M.E., Teller, S.: Automatic recovery of relative camera rotations for urban scenes. In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2000, pp. 282–289 (2000)
Košecká, J., Zhang, W.: Video compass. In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2353, pp. 476–490. Springer, Heidelberg (2002). doi:10.1007/3-540-47979-1_32
Sturm, J., Engelhard, N., Endres, F., Burgard, W., Cremers, D.: A benchmark for the evaluation of RGB-D slam systems. In: 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2012), pp. 573–580 (2012)
Kondermann, D., et al.: Stereo ground truth with error bars. In: Cremers, D., Reid, I., Saito, H., Yang, M.-H. (eds.) ACCV 2014. LNCS, vol. 9007, pp. 595–610. Springer, Cham (2015). doi:10.1007/978-3-319-16814-2_39
Geiger, A., Lenz, P., Urtasun, R.: Are we ready for autonomous driving? The Kitti vision benchmark suite. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3354–3361 (2012)
von Gioi, R., Jakubowicz, J., Morel, J.M., Randall, G.: LSD: a fast line segment detector with a false detection control. IEEE Trans. Pattern Anal. Mach. Intell. 32, 722–732 (2010)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
A Matching Test Set
A Matching Test Set
The test set consists of 8 image pairs with small baseline displacement showing different indoor and outdoor scenes. For each image in the test set, we detect line-segments using the LSD algorithmFootnote 6 [32]. Then, we manually label corresponding lines in the image pairs and save them as ground truth matches. Note that a line-segment can correspond to multiple line-segments in the other image as the line segmentation may vary.
The complete test set is shown in Fig. 14. The image pairs “Facade01” and “Facade02” are taken from the matching evaluation from Zhang et al. [13, 14], “HeiSt02” from the Heidelberger Stereo benchmark [30], “Kitti02” from the KITTI odometry benchmark [31] and“Oxford02” from the Oxford multiview datasetFootnote 7.
Image pairs of the test set.
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
von Schmude, N., Lothe, P., Witt, J., Jähne, B. (2017). Relative Pose Estimation from Straight Lines Using Optical Flow-Based Line Matching and Parallel Line Clustering. In: Braz, J., et al. Computer Vision, Imaging and Computer Graphics Theory and Applications. VISIGRAPP 2016. Communications in Computer and Information Science, vol 693. Springer, Cham. https://doi.org/10.1007/978-3-319-64870-5_16
Download citation
DOI: https://doi.org/10.1007/978-3-319-64870-5_16
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-64869-9
Online ISBN: 978-3-319-64870-5
eBook Packages: Computer ScienceComputer Science (R0)