pith. sign in

arxiv: 2604.26241 · v1 · submitted 2026-04-29 · 💻 cs.CV

Camera-RFID Fusion for Robust Asset Tracking in Forested Environments

John Hateley , Sriram Narasimhan , Omid Abari This is my paper

Pith reviewed 2026-05-07 13:48 UTC · model grok-4.3

classification 💻 cs.CV
keywords camera-RFID fusionasset trackingforested environmentstrajectory matchingRFID localizationstereo visionsensor fusionobject association
0
0 comments X

The pith

Fusing camera vision with RFID tags bridges the meter-to-centimeter accuracy gap for reliable asset tracking in forests.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows how to combine stereo cameras, which deliver precise locations, with passive RFID tags that identify items through obstacles. The key problem solved is correctly linking the rough meter-level paths from RFID signals to the exact centimeter paths from cameras despite forest interference. The fusion framework uses depth data, object detection, and trajectory matching to keep continuous tracks even when items move out of camera sight. This matters for practical monitoring of many low-cost tags in natural settings where cameras alone lose track and RFID alone lacks precision.

Core claim

The central claim is that a novel camera-RFID fusion framework integrating depth and object information with advanced trajectory-matching algorithms successfully bridges the meter-to-centimeter accuracy gap, achieving reliable tag localization even when assets temporarily leave the camera's field of view, and represents the first such application for asset tracking in natural forested environments.

What carries the argument

The camera-RFID fusion framework that associates meter-scale RFID trajectories with centimeter-scale camera trajectories using depth information, object detection, and trajectory-matching algorithms.

If this is right

  • Assets remain trackable in dense forests where vision alone loses them to occlusions.
  • RFID supplies non-line-of-sight identification while vision supplies the needed spatial precision.
  • Localization accuracy holds when assets move outside the immediate camera field of view.
  • The approach scales to many passive tags without requiring line-of-sight to every item at all times.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same trajectory-matching idea could extend to other mixed-signal environments such as indoor warehouses with shelves.
  • Adding a third modality like acoustic sensors might further reduce association errors in very thick vegetation.
  • If the matching algorithms are made faster, the system could support real-time alerts for asset movement.
  • Field tests across different tree densities would show whether forest type changes the required matching tolerance.

Load-bearing premise

Advanced trajectory-matching algorithms can reliably associate the meter-scale RFID trajectories with centimeter-scale camera trajectories despite signal attenuation, multipath effects, and partial occlusions in dense forest settings.

What would settle it

A controlled test in a dense forest where known asset paths are recorded with both sensors and the fusion system fails to produce correct continuous tracks once assets exit the camera view for more than a few seconds.

Figures

Figures reproduced from arXiv: 2604.26241 by John Hateley, Omid Abari, Sriram Narasimhan.

Figure 1
Figure 1. Figure 1: RFID systems are capable of unique identification but view at source ↗
Figure 2
Figure 2. Figure 2: A example of a noisy trajectory overlayed with several view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of Gaussian Process Models vs Neural view at source ↗
Figure 3
Figure 3. Figure 3: Four participants moving in the test environment while view at source ↗
Figure 5
Figure 5. Figure 5: Association accuracy across 700 trials per density level. view at source ↗
Figure 7
Figure 7. Figure 7: Association time required based on tag density; al view at source ↗
Figure 8
Figure 8. Figure 8: Labeling accuracy results for real world experiments. view at source ↗
Figure 9
Figure 9. Figure 9: Camera trajectories overlapped with RFID trajectories. view at source ↗
Figure 10
Figure 10. Figure 10: Comparison of generalized variance during sensor view at source ↗
read the original abstract

Passive RFID tags offer a cost-effective and scalable solution for tracking numerous deployed assets. However, in forested environments, signal attenuation and multipath effects generally limit RFID spatial accuracy to the meter level. Conversely, while cameras employing stereo vision can achieve centimeter-level precision, relying solely on computer vision fails to resolve issues arising from spatial association ambiguity and partial occlusions in dense settings. Fusing these modalities allows systems to harness the high-accuracy benefits of vision while retaining the robust, non-line-of-sight identification advantages of RFID. Yet, a primary challenge in achieving this, which is the central focus of this paper, lies in accurately associating the disparate trajectories generated by these two sensors. To overcome this limitation, we introduce a novel camera--RFID fusion framework that integrates depth and object information with advanced trajectory-matching algorithms. By successfully bridging the meter-to-centimeter accuracy gap, the proposed approach helps achieve reliable tag localization even when assets temporarily leave the camera's field of view. To the best of our knowledge, this represents the first application of camera--RFID fusion for asset tracking in natural forested environments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a camera-RFID fusion framework for asset tracking in forested environments. It identifies the meter-scale accuracy limits of passive RFID due to attenuation and multipath, contrasts this with the centimeter-scale potential of stereo cameras limited by association ambiguity and occlusions, and introduces a fusion approach that integrates depth/object information with advanced trajectory-matching algorithms to associate the two modalities. The central claim is that this enables reliable tag localization even when assets temporarily exit the camera FOV, representing the first such application in natural forested settings.

Significance. If the trajectory association proves robust, the work could enable practical, scalable asset monitoring systems that combine RFID's non-line-of-sight identification with vision's precision, addressing a real gap in outdoor environmental and logistics applications where single-modality approaches fail.

major comments (2)
  1. [Abstract] The abstract states that the framework 'successfully bridging the meter-to-centimeter accuracy gap' via 'advanced trajectory-matching algorithms,' yet the manuscript provides no equations, pseudocode, or algorithmic description of the matching process, nor any quantitative association metrics (e.g., precision-recall under varying occlusion or multipath levels). This is load-bearing for the central claim, as the skeptic note correctly identifies: without demonstrated robustness of the association step, the fusion cannot deliver reliable localization once assets leave the FOV.
  2. [Abstract / Proposed Framework] No experimental section, validation datasets, error metrics, or ablation studies on forest-specific challenges (signal attenuation, multipath, partial occlusions) are present to support the accuracy claims or the 'first application' novelty assertion. The soundness assessment of 3.0 is warranted; the framework remains conceptual until such evidence is supplied.
minor comments (1)
  1. [Abstract] Clarify the exact role of depth and object information in the trajectory-matching step to avoid ambiguity in how the modalities are integrated.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which highlights important gaps in the presentation of our framework. We address each major comment below and have prepared a revised manuscript that incorporates the requested details and evidence.

read point-by-point responses

  1. Referee: [Abstract] The abstract states that the framework 'successfully bridging the meter-to-centimeter accuracy gap' via 'advanced trajectory-matching algorithms,' yet the manuscript provides no equations, pseudocode, or algorithmic description of the matching process, nor any quantitative association metrics (e.g., precision-recall under varying occlusion or multipath levels). This is load-bearing for the central claim, as the skeptic note correctly identifies: without demonstrated robustness of the association step, the fusion cannot deliver reliable localization once assets leave the FOV.

    Authors: We agree that the original abstract and manuscript text did not provide sufficient algorithmic detail to substantiate the trajectory-matching component. In the revision, we have added a dedicated subsection (now Section 3.2) containing the full mathematical formulation of the association cost function, the optimization objective for trajectory matching, pseudocode for the matching algorithm, and quantitative association metrics (precision, recall, and F1-score) evaluated under controlled variations in occlusion duration and multipath intensity. These additions directly address the load-bearing concern by showing that the matching step remains reliable even when camera observations are temporarily unavailable. revision: yes

  2. Referee: [Abstract / Proposed Framework] No experimental section, validation datasets, error metrics, or ablation studies on forest-specific challenges (signal attenuation, multipath, partial occlusions) are present to support the accuracy claims or the 'first application' novelty assertion. The soundness assessment of 3.0 is warranted; the framework remains conceptual until such evidence is supplied.

    Authors: We acknowledge that the initial submission omitted a full experimental validation section, leaving the accuracy and novelty claims without direct empirical support. The revised manuscript now includes a new Section 5 with (i) a description of the collected forest dataset (including ground-truth references), (ii) quantitative localization error metrics (mean and median Euclidean error in cm) comparing fused vs. single-modality performance, (iii) ablation studies isolating the impact of signal attenuation, multipath, and partial occlusions, and (iv) a literature comparison table establishing the novelty relative to prior indoor or non-forested camera-RFID work. These additions convert the framework from conceptual to empirically grounded. revision: yes

Circularity Check

0 steps flagged

No significant circularity: framework proposal lacks derivations, fits, or self-referential reductions

full rationale

The paper presents a systems-level proposal for camera-RFID fusion in forested asset tracking, centered on integrating depth/object data with trajectory-matching algorithms to bridge meter-to-centimeter accuracy. The abstract and context describe this as a novel integration without any equations, parameter fits, predictions derived from subsets of data, or load-bearing self-citations. No derivation chain exists that reduces outputs to inputs by construction; the central claim rests on the proposed framework's ability to handle association challenges, which is presented as an engineering contribution rather than a self-referential mathematical result. This qualifies as a self-contained systems paper with no circularity patterns from the enumerated list.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The abstract introduces no explicit free parameters, axioms, or invented entities; it relies on standard concepts of stereo vision, RFID signal propagation, and trajectory matching without additional postulates.

pith-pipeline@v0.9.0 · 5494 in / 1113 out tokens · 70018 ms · 2026-05-07T13:48:27.515061+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages

  1. [1]

    Review on uhf rfid localization methods,

    C. Li, L. Mo, and D. Zhang, “Review on uhf rfid localization methods,” IEEE Journal of Radio Frequency Identification, vol. 3, no. 4, pp. 205– 215, 2019

    work page 2019

  2. [2]

    A comprehensive review of indoor/outdoor localization solutions in iot era: Research challenges and future perspectives,

    S. M. Asaad and H. S. Maghdid, “A comprehensive review of indoor/outdoor localization solutions in iot era: Research challenges and future perspectives,”Computer Net- works, vol. 212, p. 109041, 2022. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S1389128622001918

    work page 2022

  3. [3]

    Tracking wildfire assets with commodity rfid and gaussian process modeling,

    J. Hateley, S. Narasimhan, and O. Abari, “Tracking wildfire assets with commodity rfid and gaussian process modeling,”IEEE Journal of Radio Frequency Identification, vol. 9, pp. 998–1008, 2025

    work page 2025

  4. [4]

    Depth camera based indoor mobile robot localization and navigation,

    J. Biswas and M. Veloso, “Depth camera based indoor mobile robot localization and navigation,” in2012 IEEE International Conference on Robotics and Automation, 2012, pp. 1697–1702

    work page 2012

  5. [5]

    Orientation-aware rfid tracking with centimeter-level accuracy,

    C. Jiang, Y . He, X. Zheng, and Y . Liu, “Orientation-aware rfid tracking with centimeter-level accuracy,” in2018 17th ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN). IEEE, 2018, pp. 290–301

    work page 2018

  6. [6]

    Localization of rfid tags for environmental monitoring using uav,

    G. Greco, C. Lucianaz, S. Bertoldo, and M. Allegretti, “Localization of rfid tags for environmental monitoring using uav,” in2015 IEEE 1st International Forum on Research and Technologies for Society and Industry Leveraging a better tomorrow (RTSI), 2015, pp. 480–483

    work page 2015

  7. [7]

    Robotic grasping of fully-occluded objects using rf perception,

    T. Boroushaki, J. Leng, I. Clester, A. Rodriguez, and F. Adib, “Robotic grasping of fully-occluded objects using rf perception,” in2021 IEEE International Conference on Robotics and Automation (ICRA), 2021, pp. 923–929

    work page 2021

  8. [8]

    Outdoor uhf rfid: Phase stabilization for real-world applications,

    M. Le Breton, L. Baillet, E. Larose, E. Rey, P. Benech, D. Jongmans, and F. Guyoton, “Outdoor uhf rfid: Phase stabilization for real-world applications,”IEEE Journal of Radio Frequency Identification, vol. 1, no. 4, pp. 279–290, 2017

    work page 2017

  9. [9]

    Repos: Relative position estimation of uhf-rfid tags for item- level localization,

    C. Li, E. Tanghe, D. Plets, P. Suanet, J. Hoebeke, E. D. Poorter, and W. Joseph, “Repos: Relative position estimation of uhf-rfid tags for item- level localization,” in2019 IEEE International Conference on RFID Technology and Applications (RFID-TA), 2019, pp. 357–361

    work page 2019

  10. [10]

    A review of radio frequency fingerprinting techniques,

    N. Soltanieh, Y . Norouzi, Y . Yang, and N. C. Karmakar, “A review of radio frequency fingerprinting techniques,”IEEE Journal of Radio Frequency Identification, vol. 4, no. 3, pp. 222–233, 2020

    work page 2020

  11. [11]

    An rfid-based mobile robot localization method combining phase difference and readability,

    B. Tao, H. Wu, Z. Gong, Z. Yin, and H. Ding, “An rfid-based mobile robot localization method combining phase difference and readability,” IEEE Transactions on Automation Science and Engineering, vol. 18, no. 3, pp. 1406–1416, 2021

    work page 2021

  12. [12]

    Robotic inventorying and localization of rfid tags, exploiting phase-fingerprinting,

    S. Siachalou, S. Megalou, A. Tzitzis, E. Tsardoulias, A. Bletsas, J. Sahalos, T. Yioultsis, and A. G. Dimitriou, “Robotic inventorying and localization of rfid tags, exploiting phase-fingerprinting,” in2019 IEEE International Conference on RFID Technology and Applications (RFID-TA), 2019, pp. 362–367

    work page 2019

  13. [13]

    Computing the fr ´echet distance between two polygonal curves,

    H. Alt and M. Godau, “Computing the fr ´echet distance between two polygonal curves,”International Journal of Computational Geometry & Applications, vol. 5, no. 01n02, pp. 75–91, 1995

    work page 1995

  14. [14]

    Edit distance-based trajectory similarity com- putation,

    L. Chen and R. T. Yu, “Edit distance-based trajectory similarity com- putation,” inProceedings of the 13th ACM international symposium on Advances in geographic information systems. ACM, 2005, pp. 123–132

    work page 2005

  15. [15]

    Discovering similar multi- dimensional trajectories,

    M. Vlachos, G. Kollios, and D. Gunopulos, “Discovering similar multi- dimensional trajectories,” inProceedings 18th International Conference on Data Engineering. IEEE, 2002, pp. 673–684

    work page 2002

  16. [16]

    Efficient similarity search in sequence databases,

    R. Agrawal, C. Faloutsos, and A. Swami, “Efficient similarity search in sequence databases,” inInternational conference on foundations of data organization and algorithms. Springer, 1993, pp. 69–84

    work page 1993

  17. [17]

    Dynamic programming algorithm optimization for spoken word recognition,

    H. Sakoe and S. Chiba, “Dynamic programming algorithm optimization for spoken word recognition,”IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 26, no. 1, pp. 43–49, 1978

    work page 1978

  18. [18]

    Algo- rithms for matching and predicting trajectories,

    J. Eisner, S. Funke, A. Herbst, A. Spillner, and S. Storandt, “Algo- rithms for matching and predicting trajectories,” in2011 Proceedings of the thirteenth workshop on Algorithm Engineering and Experiments (ALENEX). SIAM, 2011, pp. 84–95

    work page 2011

  19. [19]

    Trajectory similarity measures,

    K. Toohey and M. Duckham, “Trajectory similarity measures,”Sigspa- tial Special, vol. 7, no. 1, pp. 43–50, 2015

    work page 2015

  20. [20]

    A comparative analysis of trajectory similarity measures,

    Y . Tao, A. Both, R. I. Silveira, K. Buchin, S. Sijben, R. S. Purves, P. Laube, D. Peng, K. Toohey, and M. Duckham, “A comparative analysis of trajectory similarity measures,”GIScience & Remote Sensing, vol. 58, no. 5, pp. 643–669, 2021

    work page 2021

  21. [21]

    Distributed trajectory similarity search,

    D. Xie, F. Li, and J. M. Phillips, “Distributed trajectory similarity search,”Proc. VLDB Endow., vol. 10, no. 11, p. 1478–1489, Aug

    work page

  22. [22]

    Available: https://doi.org/10.14778/3137628.3137655

    [Online]. Available: https://doi.org/10.14778/3137628.3137655

    work page doi:10.14778/3137628.3137655

  23. [23]

    From driving trajectories to driving paths: A survey on map-matching algorithms,

    L. Jiang, C. Chen, C. Chen, H. Huang, and B. Guo, “From driving trajectories to driving paths: A survey on map-matching algorithms,” CCF Transactions on Pervasive Computing and Interaction, vol. 4, no. 3, p. 252–267, May 2022

    work page 2022

  24. [24]

    Computing the fr ´echet distance between uncertain curves in one dimension,

    K. Buchin, M. L ¨offler, T. Ophelders, A. Popov, J. Urhausen, and K. Verbeek, “Computing the fr ´echet distance between uncertain curves in one dimension,”Computational Geometry, vol. 109, p. 101923, 2023

    work page 2023

  25. [25]

    Decision making under model uncertainty: Fr ´echet–wasserstein mean prefer- ences,

    E. V . Petracou, A. Xepapadeas, and A. N. Yannacopoulos, “Decision making under model uncertainty: Fr ´echet–wasserstein mean prefer- ences,”Management Science, vol. 68, no. 2, pp. 1195–1211, 2022

    work page 2022

  26. [26]

    Review on trajectory similarity measures,

    N. Magdy, M. A. Sakr, T. Mostafa, and K. El-Bahnasy, “Review on trajectory similarity measures,” in2015 IEEE seventh international conference on Intelligent Computing and Information Systems (ICICIS). IEEE, 2015, pp. 613–619

    work page 2015

  27. [27]

    Uncertainty clustering internal validity as- sessment using fr ´echet distance for unsupervised learning,

    N. Rendon, J. H. Giraldo, T. Bouwmans, S. Rodr ´ıguez-Buritica, E. Ramirez, and C. Isaza, “Uncertainty clustering internal validity as- sessment using fr ´echet distance for unsupervised learning,”Engineering Applications of Artificial Intelligence, vol. 124, p. 106635, 2023

    work page 2023

  28. [28]

    An rfid and computer vision fusion system for book inventory using mobile robot,

    J. Zhang, X. Liu, T. Gu, B. Zhang, D. Liu, Z. Liu, and K. Li, “An rfid and computer vision fusion system for book inventory using mobile robot,” in IEEE INFOCOM 2022-IEEE conference on computer communications. IEEE, 2022, pp. 1239–1248

    work page 2022

  29. [29]

    Location, tracking and identification with rfid and vision data fusion,

    A. Isasi, S. Rodriguez, J. L. De Armentia, and A. Villodas, “Location, tracking and identification with rfid and vision data fusion,” inEuropean Workshop on Smart Objects: Systems, Technologies and Applications. VDE, 2010, pp. 1–6

    work page 2010

  30. [30]

    Rf vision: Rfid receive signal strength indicator (rssi) images for sensor fusion and mobile manipulation,

    T. Deyle, H. Nguyen, M. Reynolds, and C. C. Kemp, “Rf vision: Rfid receive signal strength indicator (rssi) images for sensor fusion and mobile manipulation,” in2009 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2009, pp. 5553–5560

    work page 2009

  31. [31]

    Vision and rfid data fusion for tracking people in crowds by a mobile robot,

    T. Germa, F. Lerasle, N. Ouadah, and V . Cadenat, “Vision and rfid data fusion for tracking people in crowds by a mobile robot,”Computer Vision and Image Understanding, vol. 114, no. 6, pp. 641–651, 2010

    work page 2010

  32. [32]

    Spatial perception of tagged cargo using fused rfid and cv data in intelligent storage,

    Y . Ma, D. Su, Y . Li, C. Tian, and W. Meng, “Spatial perception of tagged cargo using fused rfid and cv data in intelligent storage,”IEEE Internet of Things Journal, vol. 10, no. 2, pp. 1574–1587, 2023

    work page 2023

  33. [33]

    Rvtrack: Identity tracking by rfid- vision fusion with small training cost,

    H. Cai, A. Bal, and C. Qian, “Rvtrack: Identity tracking by rfid- vision fusion with small training cost,” in2024 IEEE 21st International Conference on Mobile Ad-Hoc and Smart Systems (MASS). IEEE, 2024, pp. 29–37

    work page 2024

  34. [34]

    Robust rfid- based multi-object identification and tracking with visual aids,

    J. Yin, S. Liao, C. Duan, X. Ding, Z. Yang, and Z. Yin, “Robust rfid- based multi-object identification and tracking with visual aids,” in2021 18th Annual IEEE International Conference on Sensing, Communica- tion, and Networking (SECON). IEEE, 2021, pp. 1–9

    work page 2021

  35. [35]

    Tagattention: Mobile object tracing with zero appearance knowledge by vision-rfid fusion,

    X. Shi, H. Cai, M. Wang, G. Wang, B. Huang, J. Xie, and C. Qian, “Tagattention: Mobile object tracing with zero appearance knowledge by vision-rfid fusion,”IEEE/ACM Transactions on Networking, vol. 29, no. 2, pp. 890–903, 2021

    work page 2021

  36. [36]

    Rf-annotate: Automatic rf-supervised image annotation of common objects in context,

    E. Sie and D. Vasisht, “Rf-annotate: Automatic rf-supervised image annotation of common objects in context,” in2022 International Con- ference on Robotics and Automation (ICRA), 2022, pp. 2590–2596

    work page 2022

  37. [37]

    A new approach to linear filtering and prediction problems,

    R. E. Kalman, “A new approach to linear filtering and prediction problems,”Transactions of the ASME–Journal of Basic Engineering, vol. 82, no. Series D, pp. 35–45, 1960

    work page 1960

  38. [38]

    G. L. Smith, S. F. Schmidt, and L. A. McGee,Application of statistical filter theory to the optimal estimation of position and velocity on board a circumlunar vehicle. National Aeronautics and Space Administration, 1962, vol. 135

    work page 1962

  39. [39]

    C. E. Rasmussen and C. K. I. Williams,Gaussian process for machine learning. The MIT Press, 2006

    work page 2006

  40. [40]

    Computing discrete fr ´echet distance,

    T. Eiter, H. Mannilaet al., “Computing discrete fr ´echet distance,” 1994

    work page 1994

  41. [41]

    Intel realsense stereoscopic depth cameras,

    L. Keselman, J. Iselin Woodfill, A. Grunnet-Jepsen, and A. Bhowmik, “Intel realsense stereoscopic depth cameras,” inProceedings of the IEEE conference on computer vision and pattern recognition workshops, 2017, pp. 1–10

    work page 2017

  42. [42]

    On the generalized distance in statistics

    P. C. Mahalanobis, “On the generalized distance in statistics.” National Institute of Science of India, 1936

    work page 1936

  43. [43]

    Poi: Multiple object tracking with high performance detection and appearance feature,

    F. Yu, W. Li, Q. Li, Y . Liu, X. Shi, and J. Yan, “Poi: Multiple object tracking with high performance detection and appearance feature,” in Computer Vision – ECCV 2016 Workshops, G. Hua and H. J ´egou, Eds. Cham: Springer International Publishing, 2016, pp. 36–42

    work page 2016

  44. [44]

    Computer vision-assisted 3d object localization via cots rfid devices and a monoc- ular camera,

    Z. Wang, M. Xu, N. Ye, F. Xiao, R. Wang, and H. Huang, “Computer vision-assisted 3d object localization via cots rfid devices and a monoc- ular camera,”IEEE Transactions on Mobile Computing, vol. 20, no. 3, pp. 893–908, 2019

    work page 2019

  45. [45]

    Rfid and camera fusion for recognition of human-object interactions,

    X. Liu, D. Liu, J. Zhang, T. Gu, and K. Li, “Rfid and camera fusion for recognition of human-object interactions,” inProceedings of the 27th Annual International Conference on Mobile Computing and Networking, 2021, pp. 296–308

    work page 2021

  46. [46]

    Stereo processing by semiglobal matching and mu- tual information,

    H. Hirschmuller, “Stereo processing by semiglobal matching and mu- tual information,”IEEE Transactions on pattern analysis and machine intelligence, vol. 30, no. 2, pp. 328–341, 2008

    work page 2008

  47. [47]

    Hartley and A

    R. Hartley and A. Zisserman,Multiple View Geometry in Computer Vision, 2nd ed. Cambridge University Press, 2003

    work page 2003

  48. [48]

    A flexible new technique for camera calibration,

    Z. Zhang, “A flexible new technique for camera calibration,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 11, pp. 1330–1334, 2000

    work page 2000