The DOI matches its Crossref record; 62 of the 66 references it cites were found in a registry or an index.
Visual Odometry Revisited: What Should Be Learnt?
Open accessAbstract
In this work we present a monocular visual odometry (VO) algorithm which leverages geometry-based methods and deep learning. Most existing VO/SLAM systems with superior performance are based on geometry and have to be carefully designed for different application scenarios. Moreover, most monocular systems suffer from scale-drift issue. Some recent deep learning works learn VO in an end-to-end manner but the performance of these deep systems is still not comparable to geometry-based methods. In this work, we revisit the basics of VO and explore the right way for integrating deep learning with epipolar geometry and Perspective-n-Point (PnP) method. Specifically, we train two convolutional neural networks (CNNs) for estimating single-view depths and two-view optical flows as intermediate outputs. With the deep predictions, we design a simple but robust frame-to-frame VO algorithm (DF-VO) which outperforms pure deep learning-based and geometry-based methods. More importantly, our system does not suffer from the scale-drift issue being aided by a scale consistent single-view depth CNN. Extensive experiments on KITTI dataset shows the robustness of our system and a detailed ablation study shows the effect of different factors in our system. Code is available at here: DF-VO.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“As shown in (Zhan et al. 2020), DF-VO with this correspondence selection scheme has already outperformed existing VO/SLAM baselines.”
From DF-VO: What Should Be Learnt for Visual Odometry? (2021) · cites this paper as Zhan et al. 2020
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
U-Net: Convolutional Networks for Biomedical Image Segmentation
Ronneberger, O., Fischer, P., & Brox, T. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-24574-4_28 · Cited by 99,404
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Distinctive Image Features from Scale-Invariant Keypoints
Lowe, D. G. (2004) · International Journal of Computer Vision · doi:10.1023/b:visi.0000029664.99615.94 · Cited by 55,589
Multiple View Geometry in Computer Vision
Hartley, R., & Zisserman, A. (2004) · doi:10.1017/cbo9780511811685 · Cited by 21,070
Are we ready for autonomous driving? The KITTI vision benchmark suite
Geiger, A., Lenz, P., & Urtasun, R. (2012) · 2012 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2012.6248074 · Cited by 14,897
(2017) · Cited by 11,625
ORB: An efficient alternative to SIFT or SURF
Rublee, E., Rabaud, V., Konolige, K., & Bradski, G. (2011) · 2011 International Conference on Computer Vision · doi:10.1109/iccv.2011.6126544 · Cited by 11,037
Vision meets robotics: The KITTI dataset
Geiger, A., Lenz, P., Stiller, C., & Urtasun, R. (2013) · The International Journal of Robotics Research · doi:10.1177/0278364913491297 · Cited by 9,913
ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras
Mur-Artal, R., & Tardos, J. D. (2017) · IEEE Transactions on Robotics · doi:10.1109/tro.2017.2705103 · Cited by 6,898 · cited 9×
“One of the most successful and accurate full SLAM systems using a sparse (ORB) feature-based approach is ORB-SLAM2 [13], along with DSO [16], a direct keyframe-based sparse SLAM method.”
MethodInfluentialORB-SLAM: A Versatile and Accurate Monocular SLAM System
Mur-Artal, R., Montiel, J. M. M., & Tardos, J. D. (2015) · IEEE Transactions on Robotics · doi:10.1109/tro.2015.2463671 · Cited by 6,778
Research works printed but not linked: 2
- godard, 2019, Digging into self-supervised monocular depth prediction, IEEE International Conference on Computer Vision (ICCV)
- 10.1023/A:1007941100561 · DOI
Cited by
Sorted by citations
D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry
Yang, N., von Stumberg, L., Wang, R., & Cremers, D. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00136 · Cited by 438
“Recently, there have been many efforts to address this by leveraging deep neural networks [48, 68, 80, 83].”
BackgroundResultTowards Better Generalization: Joint Depth-Pose Learning Without PoseNet
Zhao, W., Liu, S., Shu, Y., & Liu, Y. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00917 · Cited by 186
“A recent paper [64] employs similar visual odometry inference strategies to utilize neural network predictions.”
MethodLocalization as a Key Enabler of 6G Wireless Systems: A Comprehensive Survey and an Outlook
Trevlakis, S. E., Boulogeorgos, A. A., Pliatsios, D., Querol, J., Ntontin, K., Sarigiannidis, P. G., Chatzinotas, S., & Renzo, M. D. (2023) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2023.3324952 · Cited by 114
Chen, C., Wang, B., Lu, C. X., Trigoni, N., & Markham, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.12567 · Cited by 145
“Zhan et al. [67] integrate the learned depth and optical flow predictions into a conventional visual odometry model, achieving competitive performance over other baselines.”
BackgroundTartanVO: A Generalizable Learning-based VOPreprint
Wang, W., Hu, Y., & Scherer, S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.00359 · Cited by 113
“Recently, Zhan et al. [29] use PnP techniques to explicitly solve for the scale factor.”
SC-DepthV3: Robust Self-Supervised Monocular Depth Estimation for Dynamic Scenes
Sun, L., Bian, J.-W., Zhan, H., Yin, W., Reid, I. D., & Shen, C. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3322549 · Cited by 119
“We focus on self-supervised learning of monocular depth from videos since only a single camera is required to collect training data in this setup, which has great potential for advancing real-world applications.”
BackgroundSelf-citationState of the Art in Vision-Based Localization Techniques for Autonomous Navigation Systems
Alkendi, Y., Seneviratne, L. D., & Zweiri, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3082778 · Cited by 114
Deep Learning for Visual Localization and Mapping: A Survey
Chen, C., Wang, B., Lu, C. X., Trigoni, N., & Markham, A. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3309809 · Cited by 97
“Furthermore, Zhan et al. [51] D3VO [17] is proposed to incorporate the predictions of depth, pose, and photometric uncertainty from DNNs into direct VO (DVO) [57].”
A Review of Sensing Technologies for Indoor Autonomous Mobile Robots
Liu, Y., Wang, S., Xie, Y., Xiong, T., & Wu, M. (2024) · Sensors · doi:10.3390/s24041222 · Cited by 102
“The ability to effectively represent the characteristics of the surrounding environment.”
BackgroundDytanVO: Joint Refinement of Visual Odometry and Motion Segmentation in Dynamic Environments
Shen, S., Cai, Y., Wang, W., & Scherer, S. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161306 · Cited by 68
“Another class of learningbased methods rely on dense optical flow to estimate pose as it provides more robust and redundant modalities for feature association in VO [5], [18], [19].”
BackgroundMonocular Visual Odometry Based on Depth and Optical Flow Using Deep Learning
Ban, X., Wang, H., Chen, T., Wang, Y., & Xiao, Y. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3024011 · Cited by 62
“A DF-VO method is proposed in [39]; it explores the optical flow and depth of per-frame extracted by DL model, combining with traditional VO to pose the localization.”
MethodAssistive Delivery Robot Application for Real-World Postal Services
Lee, D., Kang, G., Kim, B., & Shim, D. H. (2021) · IEEE Access · doi:10.1109/access.2021.3120618 · Cited by 57
“Most recent VO-related studies [19], [24]–[27] have combined vision sensors with IMU sensors for more precise position estimation; this approach is called visual inertial odometry (VIO).”
BackgroundDeep Keypoint-Based Camera Pose Estimation with Geometric Constraints
Jau, Y.-Y., Zhu, R., Su, H., & Chandraker, M. K. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341229 · Cited by 60
“For the sparse feature-based category, some work has been done using learning-based methods, e.g. [3], [32]–[34].”
MethodAdolfsson, D., Magnusson, M., Alhashimi, A., Lilienthal, A. J., & Andreasson, H. (2022) · IEEE Transactions on Robotics · doi:10.1109/tro.2022.3221302 · Cited by 54
RaLL: End-to-End Radar Localization on Lidar Map Using Differentiable Measurement Model
Yin, H., Chen, R., Wang, Y., & Xiong, R. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3061165 · Cited by 52
Real-Time Object Navigation With Deep Neural Networks and Hierarchical Reinforcement Learning
Staroverov, A., Yudin, D. A., Belkin, I., Adeshkin, V., Solomentsev, Y., & Panov, A. I. (2020) · IEEE Access · doi:10.1109/access.2020.3034524 · Cited by 53
“Modern enhancements of these approaches are neural network methods that train in a self-supervised manner – D3VO [35], DeepMatchVO [36], DF-VO [37].”
Generalizing to the Open World: Deep Visual Odometry with Online Adaptation
Li, S., Wu, X., Cao, Y., & Zha, H. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01298 · Cited by 53
SLAM integrated mobile mapping system in complex urban environments
Li, S.-X., Li, G.-Y., Wang, L., & Qin, Y. (2020) · Isprs Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2020.05.012 · Cited by 44
CFEAR Radarodometry - Conservative Filtering for Efficient and Accurate Radar Odometry
Adolfsson, D., Magnusson, M., Alhashimi, A., Lilienthal, A. J., & Andreasson, H. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9636253 · Cited by 46
Neural Network-Based Recent Research Developments in SLAM for Autonomous Ground Vehicles: A Review
Saleem, H., Malekian, R., & Munir, H. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3273913 · Cited by 45
“Moreover, Zhan et al. [68] applied an encoder–decoder CNN model with residual learning to determine the depth of monocular images for better scale estimation.”
SAGE: SLAM with Appearance and Geometry Prior for Endoscopy
Liu, X., Li, Z., Ishii, M., Hager, G. D., Taylor, R. H., & Unberath, M. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9812257 · Cited by 46
Overview of deep learning application on visual SLAM
Li, S., Zhang, D., Xian, Y., Li, B., Zhang, T., & Zhong, C. (2022) · Displays · doi:10.1016/j.displa.2022.102298 · Cited by 47
Thermal-Inertial SLAM for the Environments With Challenging Illumination
Jiang, J., Chen, X., Dai, W., Gao, Z., & Zhang, Y. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3185385 · Cited by 42
“[20] developed monocular visual odometry using deep optical flow to match two consecutive frames.”
MethodZhang, S., Zhang, J., & Tao, D. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19839-7_9 · Cited by 48
Ruhkamp, P., Gao, D., Chen, H., Navab, N., & Busam, B. (2021) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv53792.2021.00092 · Cited by 42
“The pixel-wise minimum depth error was already proposed to avoid this issue [51, 11].”
BackgroundParticleSfM: Exploiting Dense Point Trajectories for Localizing Moving Cameras in the Wild
Zhao, W., Liu, S., Guo, H., Wang, W., & Liu, Y. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19824-3_31 · Cited by 41
Spatiotemporal Registration for Event-based Visual Odometry
Liu, D., Parra, Á., & Chin, T.-J. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.00490 · Cited by 40
Autonomous Driving with Deep Learning: A Survey of State-of-Art TechnologiesPreprint
Huang, Y., & Chen, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.06091 · Cited by 84
“[256] explore a way to integrate deep learning with epipolar geometry and Perspective-n-Point (PnP) method, which trains two CNNs for estimating single-view depths and two-view optical flows as intermediate outputs.”
MethodKlenk, S., Motzet, M., Koestler, L., & Cremers, D. (2024) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv62453.2024.00036 · Cited by 42
“Drawing inspiration from the classical methods [12, 38], a beneficial approach has emerged to first predict correspondences and subsequently use these in a principled, geometry-based manner for pose prediction [52, 53, 60, 65].”
InfluentialContinual SLAM: Beyond Lifelong Simultaneous Localization and Mapping Through Continual Learning
Vödisch, N., Cattaneo, D., Burgard, W., & Valada, A. (2023) · Springer proceedings in advanced robotics · doi:10.1007/978-3-031-25555-7_3 · Cited by 43
“Other works such as DF-VO [32] eliminate the need for a pose network by leveraging feature matching based on optical flow.”
MethodEMA-VIO: Deep Visual–Inertial Odometry With External Memory Attention
Tu, Z., Chen, C., Pan, X., Liu, R., Cui, J., & Mao, J. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3208200 · Cited by 36
Li, R., Ji, P., Xu, Y., & Bhanu, B. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3207105 · Cited by 35
DF-VO: What Should Be Learnt for Visual Odometry?
Zhan, H., Weerasekera, C. S., Bian, J.-W., Garg, R., & Reid, I. (2021) · arXiv · doi:10.48550/arxiv.2103.00933 · Cited by 31
“Flow magnitude-based method (Zhan et al. 2020) is straightforward but there are some potential failure cases, which is explained in Sec.”
InfluentialSelf-citationUnsupervised Depth Learning in Challenging Indoor Video: Weak Rectification to Rescue.Preprint
Bian, J.-W., Zhan, H., Wang, N., Chin, T.-J., Shen, C., & Reid, I. D. (2020) · arXiv (Cornell University) · Cited by 26
Self-citationRadar Odometry for Autonomous Ground Vehicles: A Survey of Methods and Datasets
Abu-Alrub, N. J., & Rawashdeh, N. A. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2023.3340513 · Cited by 33
“Readers interested in deeper analysis of odometry and SLAM evaluation metrics in addition to publicly available tools for evaluation can refer to [35], [36], [37], [38], [39].”
BackgroundAccuMO: Accuracy-Centric Multitask Offloading in Edge-Assisted Mobile Augmented Reality
Kong, Z. J., Xu, Q., Meng, J., & Hu, Y. (2023) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3570361.3592531 · Cited by 24
“Secondly, the Kalman filter maintains internal states including position, velocity, orientation, and their estimated error covariance, which helps to reduce IMU sensor noise.”
BackgroundDeep Online Correction for Monocular Visual Odometry
Zhang, J., Sui, W., Wang, X., Meng, W., Zhu, H., & Zhang, Q. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561642 · Cited by 23
“Deep frame-to-frame visual odometry (DFVO) [22] utilizes CNNs to predict optical flow and depth”
InfluentialMailka, H., Abouzahir, M., & Ramzi, M. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-17595-w · Cited by 23
“…techniques, including traditional monocular and stereo approaches, and learning approaches for monocular odometry such as ORBSLAM2 [56], DF-VO [58, 59], VISO2 [57], SfM-Learner [53], SC-SfMLearner [55] and Depth-VO-Feat [54] Analyzing the results obtained, we find that the prediction model…”
SimVODIS++: Neural Semantic Visual Odometry in Dynamic Environments
Kim, U.-H., Kim, S.-H., & Kim, J.-H. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3150854 · Cited by 23
XVO: Generalized Visual Odometry via Cross-Modal Self-Training
Lai, L., Shangguan, Z., Zhang, J., & Ohn-Bar, E. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00926 · Cited by 21
“Specifically, to simplify the ill-posed monocular pose recovery task, researchers often resort to relying on knowledge of the camera intrinsic parameters to incorporate various geometric or photometric consistency-based mechanisms [37, 45, 54, 89, 112, 116, 124, 126].”
BackgroundMOLTR: Multiple Object Localization, Tracking and Reconstruction From Monocular RGB Videos
Li, K., Rezatofighi, H., & Reid, I. D. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3061080 · Cited by 20
“To this end, we demonstrate that our system can work with estimated camera trajectory from DF-VO [46] on KITTI dataset.”
BackgroundMethodInfluentialSelf-citationUnsupervised Learning of Monocular Depth and Ego-Motion in Outdoor/Indoor Environments
Gao, R., Xiao, X., Xing, W., Li, C., & Liu, L. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3151629 · Cited by 21
“To improve the accuracy of pose estimation, DF-VO [21] explores an integrating deep learning method with epipolar geometry.”
MethodNARUTO: Neural Active Reconstruction from Uncertain Target Observations
Feng, Z., Zhan, H., Chen, Z., Yan, Q., Xu, X., Cai, C., Li, B., Zhu, Q., & Xu, Y. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.02038 · Cited by 27
“These elements have led to research areas like visual odometry [60, 83], monocular depth estimation [3, 20, 23, 24, 82], multi-view stereo [6, 11, 28, 40, 63, 70, 80], structure-from-motion (SfM) [61], path planning [22, 27, 34, 35], and SLAM [5, 16, 19, 71, 75].”
BackgroundSelf-citationAutomatic Target-Less Camera-LiDAR Calibration From Motion and Deep Point Correspondences
Petek, K., Vödisch, N., Meyer, J., Cattaneo, D., Valada, A., & Burgard, W. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3468090 · Cited by 19
“In contrast to deep learning-based methods [28], [29], classical feature-based tracking approaches such as ORB-SLAM3 are more robust towards detecting lost tracks, which would lead to inconsistent constraints in the optimization problem.”
ResultZheng, J., Chen, T., He, J., Wang, Z., & Gao, B. (2024) · Energies · doi:10.3390/en17164106 · Cited by 18
Cao, Y., Zhang, X.-S., Luo, F., Peng, P., Lin, C., Yang, K.-F., & Li, Y. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.109262 · Cited by 17
“To combine the advantages of geometry-based and deep-learning methods, several works [11; 10] separately learned the various components (e.”
BackgroundMethodInfluentialDynamic Object-Aware Visual Odometry (VO) Estimation Based on Optical Flow Matching
Cho, H. M., & Kim, E. (2023) · IEEE Access · doi:10.1109/access.2023.3241961 · Cited by 17
Salient Sparse Visual Odometry With Pose-Only Supervision
Chen, S., Liu, K., Wang, C., Yuan, S., Yang, J., & Xie, L. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3384757 · Cited by 16
“Compared with end-to-end methods, the hybrid methods have better performance and generalizability [14], [21].”
BackgroundMethodInfluentialDeep Visual Odometry with Events and Frames
Pellerito, R., Cannici, M., Gehrig, D., Belhadj, J., Dubois-Matra, O., Casasco, M., & Scaramuzza, D. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801307 · Cited by 17
“Su-pervised methods, on the other hand, either rely on end-to-end camera-motion regressors [15], [16], [17], [18] or exploit hybrid solutions that combine geometric models with deep neural networks [19], [20], [21].”
Beyond Learning: Back to Geometric Essence of Visual Odometry via Fusion-Based Paradigm
Jin, Y., Wei, N., Fu, M., Liu, H., & Zhang, W. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3121483 · Cited by 16
Towards Scale Consistent Monocular Visual Odometry by Learning from the Virtual World
Zhang, S., Zhang, J., & Tao, D. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9812347 · Cited by 17
Pseudo-LiDAR for Visual Odometry
Miao, Y., Deng, H., Jiang, C., Feng, Z., Wu, X., Wang, G., & Wang, H. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3315416 · Cited by 15
Unsupervised monocular visual odometry with decoupled camera pose estimation
Lin, L., Wang, W., Luo, W., Song, L., & Zhou, W. (2021) · Digital Signal Processing · doi:10.1016/j.dsp.2021.103052 · Cited by 15
Unsupervised Visual Odometry and Action Integration for PointGoal Navigation in Indoor Environment
Cao, Y., Zhang, X.-S., Luo, F., Lin, C., & Li, Y. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3263484 · Cited by 16
“To combine the advantages of geometry-based and deep-learning methods, several works [38], [39] tried to learn the various components (e.g., optical flow, depth, and VO) of the entire system to get more accurate performance.”
Semantics Aware Dynamic SLAM Based on 3D MODT
Sualeh, M., & Kim, G.-W. (2021) · Sensors · doi:10.3390/s21196355 · Cited by 14
“The Absolute Trajectory Error (ATE) measures the root-mean-square error between predicted camera poses [x, y, z] and ground truth.”
MethodInfluentialLiDAR localization at 100 FPS: A map-aided and template descriptor-based global method
Shi, P., Li, J., & Zhang, Y. (2023) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2023.103336 · Cited by 13
Amarasinghe, C. K., Rathnaweera, A., & Maithripala, S. (2023) · Arabian Journal for Science and Engineering · doi:10.1007/s13369-023-07906-6 · Cited by 13
USD-SLAM: A Universal Visual SLAM Based on Large Segmentation Model in Dynamic Environments
Wang, J., Ren, Y., Li, Z., Xie, X., Chen, Z., Shen, T., Liu, H., & Wang, K. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3498781 · Cited by 13
“Early methods typically used geometry-based or optical flow techniques to remove dynamic objects, but these methods were computationally complex and had low accuracy [7].”
Stereo Visual Odometry Pose Correction through Unsupervised Deep Learning
Zhang, S., Lu, S., He, R., & Bao, Z. (2021) · Sensors · doi:10.3390/s21144735 · Cited by 13
“[31] presented a monocular VO system that combines the geometry-based method and the unsupervised deep learning.”
MethodGao, F., Yu, J., Shen, H., Wang, Y., & Yang, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.09369 · Cited by 15
“Zhen et al. [25] first proposed to utilize a FlowNet to predict bi-directional optical flows and samples corresponding points from flows with least forward-backward flow inconsistency to solve fundamental matrix.”
InfluentialÁlvarez-Tuñón, O., Brodskiy, Y., & Kayacan, E. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3321032 · Cited by 15
“DF-VO [61] uses a learnt optical flow for the 2-”
MethodMonocular Rotational Odometry with Incremental Rotation Averaging and Loop Closure
Chng, C.-K., Parra, Á., Chin, T.-J., & Latif, Y. (2020) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta51227.2020.9363388 · Cited by 12
UVMO: Deep unsupervised visual reconstruction-based multimodal-assisted odometry
Han, S., Li, M., Tang, H., Song, Y., & Tong, G. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110573 · Cited by 11
Wang, C., Zhang, G., Cheng, Z., & Zhou, W. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3533256 · Cited by 11
“Zhan et al. [14] proposed a hybrid VO system that leverages the advantages of two-view geometry constraint with deep learning.”
InfluentialVisual odometry errors and fault distinction for integrity monitoring
Fu, Y., Wang, S., Zhai, Y., & Zhan, X. (2020) · Aerospace Systems · doi:10.1007/s42401-020-00062-x · Cited by 11
“VO is reliable and accurate under favorable conditions, such as when there is sufficient illumination and texture to establish correspondence, sufficient overlap between consecutive frames, and when the majority of the visible scene is static [20].”
BackgroundGeometry-Constrained Scale Estimation for Monocular Visual Odometry
Zhang, H., Wang, X., Yin, X., Du, M., Liu, C., & Chen, Q. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3093771 · Cited by 13
H-VLO: Hybrid LiDAR-Camera Fusion For Self-Supervised Odometry
Aydemir, E., Fetic, N., & Ünel, M. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981111 · Cited by 11
“Similar to [7], 20% of the correspondences with the least forward-backward flow inconsistency |− F tt − 1 − F t − 1 t | are selected as correspondence matches.”
BackgroundMethodInfluentialDeep Feature-based RGB-D Odometry using SuperPoint and SuperGlue
Fujimoto, S., & Matsunaga, N. (2023) · Procedia Computer Science · doi:10.1016/j.procs.2023.10.624 · Cited by 10
Ott, F., Heublein, L., Rügamer, D., Bischl, B., & Mutschler, C. (2024) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2024.104256 · Cited by 12
“In contrast, Zhan et al. (2020) proposed the combination of a depth network with optical flow.”
BackgroundMultispectral Visual Odometry Using SVSF for Mobile Robot Localization
Fahima, B., & Nemra, A. (2021) · Unmanned Systems · doi:10.1142/s2301385022500157 · Cited by 10
Modality-invariant Visual Odometry for Embodied Vision
Memmel, M., Bachmann, R., & Zamir, A. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.02064 · Cited by 9
“To deal with the large data requirements, SLAM- and learning-based methods can be combined [5, 8, 9, 11, 48, 61, 63].”
MethodSpatiotemporal Dual-Stream Network for Visual Odometry
Xu, C., Zeng, T., Luo, Y., Song, F., & Si, B. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3544521 · Cited by 9
Visual Odometry in Challenging Environments: An Urban Underground Railway Scenario Case
Etxeberria-Garcia, M., Zamalloa, M., Arana-Arexolaleiba, N., & Labayen, M. (2022) · IEEE Access · doi:10.1109/access.2022.3187209 · Cited by 9
“Furthermore, monocular VO algorithms have a depth-translation scale ambiguity issue [30].”
BackgroundACE-DNV: Automatic classification of gaze events in dynamic natural viewing
Nejad, A., de Haan, G. A., Heutink, J., & Cornelissen, F. W. (2024) · Behavior Research Methods · doi:10.3758/s13428-024-02358-8 · Cited by 8
“Video frames are passed into the DF-VO visual odometry method to extract the camera rotation and movement, reflecting head rotation and body translation.”
MethodInfluentialActiveGAMER: Active GAussian Mapping through Efficient Rendering
Chen, L., Zhan, H., Chen, K., Xu, X., Yan, Q., Cai, C., & Xu, Y. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01537 · Cited by 8
Self-citationDecoupledPoseNet: Cascade Decoupled Pose Learning for Unsupervised Camera Ego-Motion Estimation
Zhou, W., Zhang, H., Yan, Z., Wang, W., & Lin, L. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3144958 · Cited by 8
Radar style transfer for metric robot localisation on lidar maps
Yin, H., Wang, Y., Wu, J., & Xiong, R. (2022) · CAAI Transactions on Intelligence Technology · doi:10.1049/cit2.12112 · Cited by 8
LUVI: Lightweight UWB-VIO based relative positioning for AR-IoT applications
Choi, H.-B., Lim, K.-W., & Ko, Y. (2022) · Ad Hoc Networks · doi:10.1016/j.adhoc.2023.103132 · Cited by 8
UW Deep SLAM-CNN Assisted Underwater SLAM
Amarasinghe, C. K., Ratnaweera, A., & Maitripala, S. (2023) · Applied Computer Systems · doi:10.2478/acss-2023-0010 · Cited by 8
MOFT: Monocular odometry based on deep depth and careful feature selection and tracking
Koledić, K., Cvišić, I., Marković, I., & Petrovic, I. K. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160588 · Cited by 10
MVP: Unified Motion and Visual Self-Supervised Learning for Large-Scale Robotic NavigationPreprint
Chancán, M., & Milford, M. J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.00667 · Cited by 8
“multi-sensor fusion techniques combining vision [24]–[30], inertial sensors [31]–[33], LiDAR [34]–[36], and radar [37] have been proposed to further improve both the localization accuracy and the real-time performance, with a number of recent deep-learning-based methods that can match the precision of those traditiona…”
MethodEVP-LCO: LiDAR-Camera Odometry Enhancing Vehicle Positioning for Autonomous Vehicles
Xun, Y., Dong, H., Ma, X., Mao, B., & Guo, H. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3507014 · Cited by 7
“…state-of-the-art odometry methods, such as classical LO that MULLS [38], learning-based LO that SUMA ++ [39] and LO-Net [40], classical VO that ORB-SLAM2 [21], learning-based VO that DF-VO [22], classical LCO DVL-SLAM [13], Huang et al. [41], and SDV-LOAM [42], and learning-based LCO H-VLO [43].”
Huang, R., Liu, C., Xie, H., Yu, J., Tao, T., Xu, Y., Ye, Z., & Tong, X. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3574514 · Cited by 7
“We use ORB-SLAM2, OV 2 SLAM [48], VOLDOR [49], DF-VO [50], Nope-NeRF, and LocalRF as baseline methods for comparison.”
Underwater Depth Estimation for Spherical Images
Cui, J., Jin, L., Kuang, H., Xu, Q., & Schwertfeger, S. (2021) · Journal of Robotics · doi:10.1155/2021/6644986 · Cited by 7
“investigated the end-to-end unsupervised depth-VO [39] and also integrated the depth with Perspective-n-Point (PnP) method to achieve high robustness [40].”
MethodExploiting Graph and Geodesic Distance Constraint for Deep Learning-Based Visual Odometry
Xu, F., Li, Q., Li, Q., Ding, K., & Zhu, J. (2022) · Remote Sensing · doi:10.3390/rs14081854 · Cited by 8
“To alleviate these problems, deep learning has been utilized [7–9].”
Towards Specialized Hardware for Learning-based Visual Odometry on the Edge
Chen, S., & Mai, K. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9982046 · Cited by 8
A New Dense Hybrid Stereo Visual Odometry Approach
Liu, Z., Malis, E., & Martinet, P. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981814 · Cited by 7
“When compared to the newest hybrid visual odometry method DFVO [9], which is based on sparse pose estimation solving a perspective-n-point problem, the proposed method achieve better results since we use a more accurate direct approach exploiting all the possible information in the image.”
BackgroundMethodInfluentialPose Constraints for Consistent Self-supervised Monocular Depth and Ego-Motion
Suri, Z. K. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-31438-4_23 · Cited by 7
Scale-aware direct monocular odometry
Campos, C., & Tardós, J. D. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9982173 · Cited by 7
Wang, C., Zhang, G., & Zhou, W. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610075 · Cited by 6
“Previous works either align depth to pose by supervising the depth estimation with the depth triangulated from the correspondences [6], or align pose to depth by scaling the translation with a scaling factor that is recovered by comparing the predicted depth and the triangulated depth [5].”
BackgroundMethodInfluentialWei, Y., Lu, S., Lu, W., Xiong, R., & Wang, Y. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3547696 · Cited by 6
“For instance, [7], [8], [9], [10], [11] employ supervised learning for optical flow and depth estimation, integrating multi-view geometry and backend optimization for improved results.”
InfluentialDINO-VO: A Feature-Based Visual Odometry Leveraging a Visual Foundation Model
Azhari, M. B., & Shim, D. H. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3592065 · Cited by 9
“Despite never having seen these scenarios, DINO-VO provides significant improvement against general frame-to-frame VO methods [7], [23], [24], [37] and even methods which are designed/trained for driving scenarios [5], [25], [33], [38], [39], [40], showcasing its generalization ability.”
Road-Network-Map-Assisted Vehicle Positioning Based on Pose Graph Optimization
Xu, S., Sun, Y., Zhao, K., Fu, X., & Wang, S. (2023) · Sensors · doi:10.3390/s23177581 · Cited by 6
Adaptive Learning for Hybrid Visual Odometry
Liu, Z., Malis, E., & Martinet, P. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3418271 · Cited by 6
Mamba-VNPS: A Visual Navigation and Positioning System with State-Selection Space
Huang, L., Wang, Z., Xiong, Q., Qu, R., Yao, C., & Li, C. (2024) · Drones · doi:10.3390/drones8110663 · Cited by 5
Geometry-Constrained Monocular Scale Estimation Using Semantic Segmentation for Dynamic Scenes
Zhang, H., Wu, Z., Li, H., Shangguan, Q., & An, K. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3625332 · Cited by 5
“Its primary objective is to estimate ego motion from consecutive robot visual signals [3], [4].”
SuperVO: A Monocular Visual Odometry based on Learned Feature Matching with GNN
Rao, S. (2021) · International Conference on Consumer Electronics and Computer Engineering (ICCECE) · doi:10.1109/iccece51280.2021.9342136 · Cited by 5
“Th e c o m p a r i s o n w i t h o t h e r m e t h o d 's ATE, p a r t o f t h e D A TA CA M E FR O M [16]”
BackgroundUrban localization using robust filtering at multiple linearization points
Gupta, S., Mohanty, A. D., & Gao, G. (2023) · EURASIP Journal on Advances in Signal Processing · doi:10.1186/s13634-023-01062-7 · Cited by 5
“This strategy, which combines monocular depth estimates with keypoints detected on the image, has been previously explored in [65, 66] and has been shown to outperform both geometry-based and learning-based methods.”
BackgroundLiLO: lightweight and low-bias LiDAR odometry method based on spherical range image filtering
Velasco-Sánchez, E., Muñoz-Bañón, M. Á., Candelas-Herías, F. A., Méndez, S. T. P., & Torres, F. (2025) · Advanced Robotics · doi:10.1080/01691864.2025.2571569 · Cited by 5
“Also, we verified the results of translation and rotation error rates using the evaluation tool presented in [42], where the average translation error (ATE) and average rotation error (ARE) are calculated and defined in [43].”
Monocular visual odometry for underground railway scenarios
Etxeberria-Garcia, M., Labayen, M., Eizaguirre, F., Zamalloa, M., & Arana-Arexolaleiba, N. (2021) · International Conference on Quality Control by Artificial Vision (QCAV) · doi:10.1117/12.2586310 · Cited by 5
“presented DF-VO.(18) It outperforms pure deep learning-based and geometry-based methods and solves the scale-drift issue by adding a scale consistent single-view depth CNN.”
Background
Metadata
- Authors
- Zhan, H., Weerasekera, C. S., Bian, J.-W., & Reid, I.
- Affiliations
- The University of Adelaide (AU)
- Venue
- 2020 IEEE International Conference on Robotics and Automation (ICRA)
- Publisher
- IEEE
- Published
- 2020-05
- Pages
- 4203-4210
- ISSN
- 1050-4729 (ISSN-L)
- OpenAlex
- W3091687063
- Cited by
- 201
- References
- 66
- Citations by year
- 2026: 18 · 2025: 28 · 2024: 37 · 2023: 37 · 2022: 33 · 2021: 30 · 2020: 17
- FWCI
- 43.12
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- Proceedings - IEEE International Conference on Robotics and Automation/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Adelaide Research & Scholarship (AR&S) (University of Adelaide) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Australian Centre for Robotic Vision
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Robotics and Sensor-Based Localization
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Physical Sciences › Earth and Planetary Sciences › Geology › 3D Surveying and Cultural Heritage
- Keywords
- monocular visual odometry, scale drift, single-view depth estimation, optical flow, epipolar geometry, perspective-n-point, convolutional neural networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
