SPATIAL-GEOMETRIC EVALUATION OF LOCAL FEATURES IN MONOCULAR VISUAL ODOMETRY

Andriy Fesiuk, Yuriy Furgala

Abstract


Background. Monocular visual odometry is an important component of visual navigation systems. However, its accuracy depends on the quality of local features and inter-frame correspondences. In the VO task, not only is geometric consistency important, but also motion observability, the physical validity of the recovered configuration, and the spatial-structural properties of local features. This study aims to provide a comprehensive evaluation of keypoint detection and description methods for monocular visual odometry.

Materials and Methods. The study was conducted on the EuRoC MAV dataset. The ORB, BRISK, AKAZE, KAZE, SIFT, SURF, and SuperPoint methods were analyzed for the number of keypoints, ranging from 200 to 1000. Motion estimation was performed using the essential matrix, the USAC_FAST filter, the recoverPose method, a minimum parallax check, and spatially guided keypoint selection. The accuracy of the recovered trajectory was evaluated using the APE and RPE metrics. To analyze the quality of local features and correspondences, the geometric component, the parallax indicator, the correct cheirality ratio, and metrics of keypoint coverage uniformity, local redundancy, and structural consistency were used. An integral quality indicator was applied to summarize the results.

Results and Discussion. The geometric metrics most often highlight AKAZE and SURF, whereas SuperPoint shows strong performance in terms of spatial characteristics. In terms of the structural consistency of correspondences, SURF consistently demonstrates the best results. As the number of keypoints increases, most methods show an initial improvement followed by saturation, and in some cases, a deterioration of individual characteristics. SURF was found to be the most balanced method across the set of criteria, whereas ORB showed the weakest results in most cases. The correlation analysis showed that the informativeness of the metrics varies by sequence type.

Conclusion. The proposed approach confirmed the relevance of multicriteria evaluation of local features in monocular visual odometry. It was shown that no single metric is universal across all scene types. In contrast, the integral indicator enables the summary of different aspects of quality and a more well-grounded ranking of the methods.

Keywords: monocular visual odometry, keypoint detection, image matching, motion estimation, deep learning, neural networks.


Full Text:

PDF

References


[1] Herrera-Granda, E. P., Torres-Cantero, A., Haro, G., & Nieto, M. (2024). Monocular visual SLAM, visual odometry, and structure from motion methods applied to 3D reconstruction: A comprehensive survey. Heliyon, 10(18), e37356. https://doi.org/10.1016/j.heliyon.2024.e37356

[2] Zhao, L., Li, Y., Wang, M., & Zhang, Y. (2025). PLL-VO: An efficient and robust visual odometry integrating point-line features and neural networks. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, X-G-2025, 1045-1052. https://doi.org/10.5194/isprs-annals-X-G-2025-1045-2025

[3] Luan, S., Yang, C., Qin, X., Chen, D., & Sui, W. (2024). Towards robust visual odometry by motion blur recovery. Frontiers in Signal Processing, 4, 1417363. https://doi.org/10.3389/frsip.2024.1417363

[4] Budzan, S., Wyżgolik, R., & Lysko, M. (2025). Performance analysis of keypoints detection and description algorithms for stereo vision based odometry. Sensors, 25(19), 6129.https://doi.org/10.3390/s25196129

[5] Huang, Q., Guo, X., Wang, Y., Sun, H., & Yang, L. (2024). A survey of feature matching methods. IET Image Processing, 18(6), 1385-1410. https://doi.org/10.1049/ipr2.13032

[6] Fesiuk, A., & Furgala, Y. (2025). Comprehensive spatial-geometric evaluation of keypoint detectors. Electronics and Information Technologies, 32, 67-86. https://doi.org/10.30970/eli.32.5

[7] Decker, P., Paulus, D., & Feldmann, T. (2008). Dealing with degeneracy in essential matrix estimation. In 2008 15th IEEE International Conference on Image Processing (pp. 1964-1967). IEEE. https://doi.org/10.1109/ICIP.2008.4712167

[8] Ma, J., Jiang, X., Jiang, J., Zhao, J., & Guo, X. (2021). Image matching from handcrafted to deep features: A survey. International Journal of Computer Vision, 129(1), 23-79. https://doi.org/10.1007/s11263-020-01359-2

[9] Xu, S., Chen, S., Xu, R., Wang, C., Lu, P., & Guo, L. (2024). Local feature matching using deep learning: A survey. Information Fusion, 107, 102344. https://doi.org/10.1016/j.inffus.2024.102344

[10] Nagy, A., Barsi, Á., & Takács, B. (2025). A comparative evaluation of classical and deep learning visual odometry configurations. Engineering Proceedings, 113(1), 16. https://doi.org/10.3390/engproc2025113016

[11] Yu, J., Zhang, H., Liu, Q., & Chen, Z. (2025). Dynamic feature elimination-based visual–inertial odometry based on an optimized SuperPoint feature extractor. Sensors, 26(1), 52. https://doi.org/10.3390/s26010052

[12] Lowe, D. G. (2004). Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2), 91–110. https://doi.org/10.1023/B:VISI.0000029664.99615.94

[13] Bay, H., Ess, A., Tuytelaars, T., & Van Gool, L. (2008). Speeded-Up Robust Features (SURF). Computer Vision and Image Understanding, 110(3), 346–359. https://doi.org/10.1016/j.cviu.2007.09.014

[14] Alcantarilla, P. F., Bartoli, A., & Davison, A. J. (2012). KAZE features. In ECCV 2012 (LNCS 7577, pp. 214–227). https://doi.org/10.1007/978-3-642-33783-3_16

[15] Alcantarilla, P. F., Nuevo, J., & Bartoli, A. (2013). Fast explicit diffusion for accelerated features in nonlinear scale spaces. In BMVC 2013 (pp. 1–11). https://doi.org/10.5244/C.27.13

[16] Rublee, E., Rabaud, V., Konolige, K., & Bradski, G. (2011). ORB: An efficient alternative to SIFT or SURF. In ICCV 2011 (pp. 2564–2571). https://doi.org/10.1109/ICCV.2011.6126544

[17] Leutenegger, S., Chli, M., & Siegwart, R. Y. (2011). BRISK: Binary robust invariant scalable keypoints. In ICCV 2011 (pp. 2548–2555). https://doi.org/10.1109/ICCV.2011.6126542

[18] DeTone, D., Malisiewicz, T., & Rabinovich, A. (2018). SuperPoint: Self-supervised interest point detection and description. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (pp. 224-236). https://doi.org/10.48550/arXiv.1712.07629

[19] Budzan, S., Wyżgolik, R., & Lysko, M. (2025). Performance analysis of keypoints detection and description algorithms for stereo vision based odometry. Sensors, 25(19), 6129. https://doi.org/10.3390/s25196129

[20] Burri, M., Nikolic, J., Gohl, P., Schneider, T., Rehder, J., Omari, S., Achtelik, M. W., & Siegwart, R. (2016). The EuRoC micro aerial vehicle datasets. The International Journal of Robotics Research, 35(10), 1157-1163. https://doi.org/10.1177/0278364915620033

[21] Lindenberger, P., Sarlin, P.-E., & Pollefeys, M. (2023). LightGlue: Local Feature Matching at Light Speed. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 17681-17692).https://doi.org/10.1109/ICCV51070.2023.01616

[22] Nannen, V., de Rezende, P. J., & Oliveira, M. M. (2013). Grid-based spatial keypoint selection for real time visual odometry. In Proceedings of the International Conference on Computer Vision Theory and Applications (VISAPP) (pp. 1-8).https://doi.org/10.5220/0004270005860589

[23] Cho, H. M., Lee, S. H., & Kim, H. S. (2025). Robust visual–inertial odometry via multi-scale deep feature extraction and adaptive keypoint selection. Applied Sciences, 15(20), 10935. https://doi.org/10.3390/app152010935

[24] Howse, Joseph, and Joe Minichino. “Learning OpenCV 4 Computer Vision with Python 3: Get to grips with tools, techniques, and algorithms for computer vision and machine learning.”, Packt Publishing Ltd, 2020.

[25] Fesiuk, A., & Furgala, Y. (2025). Keypoint matches filtering in computer vision: Comparative analysis of RANSAC and USAC variants. International Journal of Computing, 24(2), 343–350. https://doi.org/10.47839/ijc.24.2.4018

[26] M. Ivashechkin, D. Baráth, J. Matas, "USACv20: Robust Essential, Fundamental and Homography Matrix Estimation," 2021. https://doi.org/10.48550/arXiv.2104.05044

[27] Nistér, D. (2004). An efficient solution to the five-point relative pose problem. IEEE Transactions on Pattern Analysis and Machine Intelligence, 26(6), 756-770. https://doi.org/10.1109/TPAMI.2004.17

[28] Grupp, M. (2026). evo: Python package for the evaluation of odometry and SLAM. GitHub repository. https://github.com/MichaelGrupp/evo

[29] Umeyama, S. (1991). Least-squares estimation of transformation parameters between two point patterns. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13(4), 376-380. https://doi.org/10.1109/34.88573

[30] Hartley, R., & Zisserman, A. (2004). Multiple View Geometry in Computer Vision (2nd ed.). Cambridge University Press.

[31] Hossein-Nejad, Z., & Nasri, M. (2017). RKEM: Redundant keypoint elimination method in image registration. IET Image Processing, 11(4), 273-284. https://doi.org/10.1049/iet-ipr.2016.0440

[32] Bailo, O., Rameau, F., Joo, K., Park, J., Bogdan, O., & Kweon, I. S. (2018). Efficient adaptive non-maximal suppression algorithms for homogeneous spatial keypoint distribution. Pattern Recognition Letters, 106, 53-60. https://doi.org/10.1016/j.patrec.2018.02.020

[33] Gauglitz, S., Foschini, L., Turk, M., & Höllerer, T. (2011). Efficiently selecting spatially distributed keypoints for visual tracking. In 2011 18th IEEE International Conference on Image Processing (pp. 1869-1872). IEEE. https://doi.org/10.1109/ICIP.2011.6115832




DOI: http://dx.doi.org/10.30970/eli.33.9

Refbacks

  • There are currently no refbacks.