pith. sign in

arxiv: 2605.15779 · v1 · pith:ICMJ7YCDnew · submitted 2026-05-15 · 💻 cs.RO · cs.AI

A Topology-Aware Spatiotemporal Handover Framework for Continuous Multi-UAV Tracking

Jianlin Ye , Christos Kyrkou , Panayiotis Kolios This is my paper

Pith reviewed 2026-05-20 18:41 UTC · model grok-4.3

classification 💻 cs.RO cs.AI
keywords multi-UAV trackingvehicle trackinghandover mechanismtopology-awarespatiotemporalmulti-camera multi-vehicle trackingintelligent transportation systemsidentity persistence
0
0 comments X

The pith

A queue-based algorithm using geometric overlaps and virtual lanes maintains vehicle identities across multiple UAV views for continuous tracking.

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

The paper aims to solve the problem of vehicle identity loss when tracking moves from one UAV to another in a network. By relying on the spatial arrangement of camera views and simplified lane models rather than visual appearance, it proposes a way to hand over tracking responsibilities predictively. This matters because fragmented trajectories prevent larger-scale analysis of traffic patterns such as where vehicles start and end their journeys. If successful, the method allows multiple drones to act as a single coordinated sensor network for intelligent transportation systems.

Core claim

The central claim is that a deterministic queue-based matching algorithm, which uses geometric overlaps between UAV fields of view and virtual lane discretization, can predictively manage identity handovers via FIFO queues. This approach achieves reliable global identity persistence in multi-UAV setups for vehicle tracking in urban environments without depending on appearance-based re-identification.

What carries the argument

The deterministic queue-based matching algorithm that utilizes geometric overlaps and virtual lane discretization to manage identity handover through FIFO queues.

If this is right

  • Vehicle trajectories remain continuous across multiple UAVs, enabling network-level traffic analysis.
  • Real-time processing of multiple 4K video streams becomes feasible for edge deployment.
  • The system outperforms traditional re-identification methods in handover accuracy for complex traffic scenarios like intersections.
  • Scalable multi-UAV deployment for traffic monitoring is supported without heavy computational costs.

Where Pith is reading between the lines

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

  • Similar queue mechanisms could apply to other multi-agent tracking problems where spatial topology is known.
  • Integration with ground-based sensors might further improve robustness in areas with poor UAV coverage.
  • Testing in varying weather conditions could reveal limits of relying solely on geometric information.

Load-bearing premise

The method assumes that geometric overlaps between UAV fields of view and virtual lane discretization give enough reliable information to handle identity handovers accurately without any appearance-based features.

What would settle it

A sequence of vehicle paths through an intersection or merging lane where multiple UAV views overlap but the queue-based system incorrectly switches or loses identities would indicate the claim does not hold.

Figures

Figures reproduced from arXiv: 2605.15779 by Christos Kyrkou, Jianlin Ye, Panayiotis Kolios.

Figure 1
Figure 1. Figure 1: The remainder of this paper is organized as follows. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 1
Figure 1. Figure 1: The proposed Unified Aerial Surveillance (UAS) framework for multi-UAV urban vehicle tracking. The pipeline is structured into three hierarchical [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Single-UAV processing outputs and key appearance challenges in nadir-view aerial traffic monitoring. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Experimental data acquisition [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Accuracy-efficiency trade-off analysis on the VisDrone2019-val [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Inference Latency vs. Accuracy for YOLO11 variants on Jetson [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Operational snapshot of the proposed Unified Aerial Surveillance system. Three synchronized UAV video streams [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
read the original abstract

The integration of Unmanned Aerial Vehicles(UAVs) into Intelligent Transportation Systems (ITS) offers synoptic visibility for traffic monitoring, yet scalable deployment is hindered by trajectory fragmentation, where vehicle identity persistence is lost across multi-UAV Fields of View (FOV). While state-of-the-art frameworks excel in optimizing local trajectory extraction and stability for single-drone imagery, they often function as isolated data silos that generate disjointed trajectories, thereby precluding network-level analysis such as Origin-Destination estimation. This paper presents a real-time Multi-Camera Multi-Vehicle Tracking (MCMT) system designed to handle global identity persistence. Addressing the visual ambiguity and computational cost of appearance-based Re-Identification (Re-ID) in nadir views, we introduce a lightweight Topology-Based Spatiotemporal Handover mechanism. We implement a high-throughput parallel pipeline leveraging YOLO11 and ByteTrack to process concurrent 4K streams. Our core contribution is a deterministic queue-based matching algorithm that utilizes geometric overlaps and virtual lane discretization to predictively manage identity handover via FIFO queues. Experimental results on complex urban environments, including intersections and merging traffic, demonstrate a Handover Success Rate (HOSR) of 99.8% in continuous traffic flows, significantly outperforming Re-ID baselines (74.1%) while validating edge deployment feasibility. The source code is available at https://github.com/JYe9/multi-camera-multi-vehicle-tracking-system.

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

1 major / 1 minor

Summary. The paper presents a real-time Multi-Camera Multi-Vehicle Tracking (MCMT) system for continuous multi-UAV vehicle tracking in intelligent transportation systems. It introduces a Topology-Based Spatiotemporal Handover mechanism consisting of a deterministic queue-based matching algorithm that uses geometric overlaps between UAV fields of view and virtual lane discretization to manage identity handovers through FIFO queues. The pipeline processes concurrent 4K streams using YOLO11 and ByteTrack, and experimental results on complex urban environments (including intersections and merging traffic) report a Handover Success Rate (HOSR) of 99.8%, substantially outperforming appearance-based Re-ID baselines at 74.1%. The source code is made available.

Significance. If the central performance claims hold under broader validation, the work provides a lightweight, geometry-driven alternative to Re-ID for maintaining trajectory continuity across multiple UAVs, which could enable network-level analyses such as origin-destination estimation in ITS. The deterministic, parameter-free character of the queue-based algorithm and the public release of the implementation code are clear strengths supporting reproducibility and edge deployment.

major comments (1)
  1. [Experimental results] Experimental results section: the aggregate HOSR of 99.8% is reported over scenes that include merging traffic, yet no breakdown, ablation, or isolated error analysis is provided for lane-changing events. Such maneuvers would violate the fixed virtual lane discretization and FIFO ordering assumptions that underpin the deterministic handover logic, rendering the headline figure sensitive to the (unstated) proportion of lane-keeping trajectories in the test data.
minor comments (1)
  1. [Method] The description of virtual lane discretization would benefit from an explicit definition or diagram early in the method section to clarify how lanes are constructed from the geometric overlaps.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment on our experimental evaluation. We address the point directly below and commit to revisions that strengthen the presentation of results.

read point-by-point responses

  1. Referee: [Experimental results] Experimental results section: the aggregate HOSR of 99.8% is reported over scenes that include merging traffic, yet no breakdown, ablation, or isolated error analysis is provided for lane-changing events. Such maneuvers would violate the fixed virtual lane discretization and FIFO ordering assumptions that underpin the deterministic handover logic, rendering the headline figure sensitive to the (unstated) proportion of lane-keeping trajectories in the test data.

    Authors: We agree that an isolated analysis of lane-changing events would improve transparency. The reported 99.8% HOSR was measured across all trajectories in the test scenes, which explicitly include merging traffic and intersections where lane changes occur. The virtual lane discretization is derived from scene topology rather than being rigidly fixed; geometric overlap detection allows dynamic queue reassignment when a vehicle crosses lane boundaries, preserving FIFO ordering only within each active segment. Nevertheless, the original manuscript does not provide a per-maneuver breakdown or ablation isolating lane-change cases. We will add this analysis, including error rates on lane-changing subsets, to the revised Experimental Results section. revision: yes

Circularity Check

0 steps flagged

No significant circularity in deterministic geometric algorithm

full rationale

The paper's core contribution is a deterministic queue-based matching algorithm that uses geometric overlaps and virtual lane discretization to manage identity handovers via FIFO queues. This is an explicit algorithmic construction based on topology and geometry rather than any fitted parameters, self-definitions, or predictions that reduce to the inputs by construction. The reported HOSR of 99.8% is presented as an experimental result on urban test scenes, not a derived claim that loops back to the method's assumptions. No self-citations are invoked as load-bearing uniqueness theorems, and the approach does not rename known results or smuggle ansatzes. The derivation chain is self-contained as a rule-based system whose correctness is evaluated externally against Re-ID baselines.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Based solely on the abstract, the system rests on domain assumptions about camera geometry and traffic modeling rather than new free parameters or invented entities. No fitted constants or novel physical entities are mentioned.

axioms (2)
  • domain assumption UAV fields of view have sufficient geometric overlaps to enable reliable predictive handover
    Invoked as the basis for the queue-based matching algorithm in the core contribution description.
  • domain assumption Virtual lane discretization accurately captures vehicle movement patterns in urban traffic
    Used to support identity management in the spatiotemporal handover mechanism.

pith-pipeline@v0.9.0 · 5792 in / 1499 out tokens · 81846 ms · 2026-05-20T18:41:35.084994+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

20 extracted references · 20 canonical work pages · 1 internal anchor

  1. [1]

    A review of computer vision techniques for the analysis of urban traffic,

    N. Buch, S. A. Velastin, and J. Orwell, “A review of computer vision techniques for the analysis of urban traffic,”IEEE Transactions on Intelligent Transportation Systems, vol. 12, no. 3, pp. 920–939, 2011

    work page 2011

  2. [2]

    A survey of unmanned aerial vehicles (uavs) for traffic monitoring,

    K. Kanistras, G. Martins, M. J. Rutherford, and K. P. Valavanis, “A survey of unmanned aerial vehicles (uavs) for traffic monitoring,” in 2013 international conference on unmanned aircraft systems (ICUAS). IEEE, 2013, pp. 221–234

    work page 2013

  3. [3]

    Unmanned aerial aircraft systems for transportation engineering: Current practice and future challenges,

    E. N. Barmpounakis, E. I. Vlahogianni, and J. C. Golias, “Unmanned aerial aircraft systems for transportation engineering: Current practice and future challenges,”International Journal of Transportation Sci- ence and Technology, vol. 5, no. 3, pp. 111–122, 2016

    work page 2016

  4. [4]

    A systematic review of drone based road traffic monitoring system,

    I. Bisio, C. Garibotto, H. Haleem, F. Lavagetto, and A. Sciarrone, “A systematic review of drone based road traffic monitoring system,” IEEE Access, vol. 10, pp. 101 537–101 555, 2022

    work page 2022

  5. [5]

    The highD dataset: A drone dataset of naturalistic vehicle trajectories on german highways for validation of highly automated driving systems,

    R. Krajewski, J. Bock, L. Kloeker, and L. Eckstein, “The highD dataset: A drone dataset of naturalistic vehicle trajectories on german highways for validation of highly automated driving systems,” in2018 21st International Conference on Intelligent Transportation Systems (ITSC). IEEE, 2018, pp. 2118–2125

    work page 2018

  6. [6]

    Cityflow: A city-scale benchmark for multi-target multi-camera vehicle tracking and re- identification,

    Z. Tang, M. Naphade, M.-Y . Liu, X. Yang, S. Birchfield, S. Wang, R. Kumar, D. Anastasiu, and J.-N. Hwang, “Cityflow: A city-scale benchmark for multi-target multi-camera vehicle tracking and re- identification,” inProceedings of the IEEE/CVF conference on com- puter vision and pattern recognition, 2019, pp. 8797–8806

    work page 2019

  7. [7]

    Visdrone-det2019: The vision meets drone object detection in image challenge results,

    D. Du, P. Zhu, L. Wen, X. Bian, H. Lin, Q. Hu, T. Peng, J. Zheng, X. Wang, Y . Zhanget al., “Visdrone-det2019: The vision meets drone object detection in image challenge results,” inProceedings of the IEEE/CVF international conference on computer vision workshops, 2019, pp. 0–0

    work page 2019

  8. [8]

    Dota: A large-scale dataset for object detection in aerial images,

    G.-S. Xia, X. Bai, J. Ding, Z. Zhu, S. Belongie, J. Luo, M. Datcu, M. Pelillo, and L. Zhang, “Dota: A large-scale dataset for object detection in aerial images,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 3974–3983

    work page 2018

  9. [9]

    You only look once: Unified, real-time object detection,

    J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788

    work page 2016

  10. [10]

    YOLOX: Exceeding YOLO Series in 2021

    Z. Ge, S. Liu, F. Wang, Z. Li, and J. Sun, “Yolox: Exceeding yolo series in 2021,”arXiv preprint arXiv:2107.08430, 2021

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  11. [11]

    Ultralytics yolo11,

    G. Jocher and J. Qiu, “Ultralytics yolo11,” 2024. [Online]. Available: https://github.com/ultralytics/ultralytics

    work page 2024

  12. [12]

    Bytetrack: Multi-object tracking by associating every detection box,

    Y . Zhang, P. Sun, Y . Jiang, D. Yu, F. Weng, Z. Yuan, P. Luo, W. Liu, and X. Wang, “Bytetrack: Multi-object tracking by associating every detection box,” inEuropean conference on computer vision. Springer, 2022, pp. 1–21

    work page 2022

  13. [13]

    Observation- centric sort: Rethinking sort for robust multi-object tracking,

    J. Cao, J. Pang, X. Weng, R. Khirodkar, and K. Kitani, “Observation- centric sort: Rethinking sort for robust multi-object tracking,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 9686–9696

    work page 2023

  14. [14]

    Multi-camera tracking of vehicles based on deep features re-id and trajectory-based camera link models

    H.-M. Hsu, T.-W. Huang, G. Wang, J. Cai, Z. Lei, and J.-N. Hwang, “Multi-camera tracking of vehicles based on deep features re-id and trajectory-based camera link models.” inCVPR workshops, 2019, pp. 416–424

    work page 2019

  15. [15]

    Multi- target multi-camera tracking of vehicles using metadata-aided re-id and trajectory-based camera link model,

    H.-M. Hsu, J. Cai, Y . Wang, J.-N. Hwang, and K.-J. Kim, “Multi- target multi-camera tracking of vehicles using metadata-aided re-id and trajectory-based camera link model,”IEEE Transactions on Image Processing, vol. 30, pp. 5198–5210, 2021

    work page 2021

  16. [16]

    Distributed multiple model mpc for target tracking uavs,

    S. Wolfe, S. Givigi, and C.-A. Rabbath, “Distributed multiple model mpc for target tracking uavs,” in2020 International Conference on Unmanned Aircraft Systems (ICUAS). IEEE, 2020, pp. 123–130

    work page 2020

  17. [17]

    Cooperative multi-uav system for surveillance and search&rescue operations over a mobile 5g node,

    R. Zahinos, H. Abaunza, J. Murillo, M. Trujillo, and A. Viguria, “Cooperative multi-uav system for surveillance and search&rescue operations over a mobile 5g node,” in2022 International Conference on Unmanned Aircraft Systems (ICUAS). IEEE, 2022, pp. 1016–1024

    work page 2022

  18. [18]

    Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery,

    R. Fonod, H. Cho, H. Yeo, and N. Geroliminis, “Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery,”Transportation Research Part C: Emerging Technologies, vol. 178, p. 105205, 2025

    work page 2025

  19. [19]

    Deep cosine metric learning for person re-identification,

    N. Wojke and A. Bewley, “Deep cosine metric learning for person re-identification,” in2018 IEEE Winter Conference on Applications of Computer Vision (WACV). IEEE, 2018, pp. 748–756

    work page 2018

  20. [20]

    Fastreid: A pytorch toolbox for general instance re-identification,

    L. He, X. Liao, W. Liu, X. Liu, P. Cheng, and T. Mei, “Fastreid: A pytorch toolbox for general instance re-identification,” inProceedings of the 31st ACM International Conference on Multimedia, 2023, pp. 9664–9667

    work page 2023