Back

A Benchmark and Simulator for UAV Tracking

Matthias MuellerNeil G. SmithBernard Ghanem

article2016ECCV1,890 citations

Presents a dedicated low-altitude aerial tracking benchmark of 123 fully annotated HD video sequences alongside an Unreal Engine-based simulator for real-time evaluation and synthetic data generation.

Listen

Deploying automated visual tracking on unmanned aerial vehicles (UAVs) offers significant utility across applications such as search and rescue, surveillance, wildlife monitoring, and navigation. However, existing tracking algorithms have primarily been evaluated on ground-based datasets that fail to capture the specific conditions encountered during airborne flight, such as abrupt camera motion, rapid aspect ratio shifts, and drastic scale changes.

The article establishes a dedicated baseline for low-altitude aerial tracking by evaluating state-of-the-art computer vision trackers against a new high-definition aerial dataset and introducing a photo-realistic, closed-loop UAV simulator for real-time evaluation.

To achieve this, the authors curated the UAV123 dataset, comprising 123 fully annotated video sequences and over 110,000 frames captured from low-altitude flights, small low-cost drones, and synthetic sequences. They benchmarked 14 leading tracking algorithms using standard precision and success metrics across diverse visual attributes. In addition, they developed an Unreal Engine 4 simulator to evaluate how tracking latency and accuracy directly interact with UAV flight dynamics and control systems in real time.

The investigation produced several key findings: First, trackers performed substantially worse on aerial video than on standard ground-based benchmarks, exposing distinct vulnerabilities to scale variation, aspect ratio changes, and low resolution. Second, correlation-filter-based algorithms incorporating scale adaptation, specifically SRDCF, achieved top performance on the overall aerial benchmark, while algorithms with memory-retention mechanisms like MUSTER excelled in long-term tracking scenarios. Third, tracker processing speed emerged as a critical operational bottleneck; when video processing dropped to 10 frames per second to emulate compute constraints, tracker accuracy degraded by 7% to 36%. Finally, closed-loop simulation revealed that tracker lag creates instability during rapid maneuvers, though increasing UAV flight altitude or fine-tuning flight controllers helped mitigate tracking errors caused by computational delay.

These results demonstrate that standard computer vision trackers cannot be directly deployed onto aerial platforms without substantial performance degradation. For operational systems, algorithm latency introduces physical safety risks and mission failure when drones attempt to follow dynamic targets, meaning speed is just as critical as baseline accuracy.

For future development and system deployment, engineering teams should prioritize integrating scale and aspect-ratio adaptation into lightweight, correlation-based trackers. System designers should utilize closed-loop simulation environments to validate tracking algorithms against flight control parameters before field testing, while considering dynamic flight tactics—such as temporarily increasing altitude—to maintain target locks when tracking fast-moving objects.

While the benchmark and simulator provide robust, highly controlled testing environments, the evaluations rely on standardized parameter sets and a specific drone model. Readers should note that performance may vary under harsher field conditions, diverse sensor payloads, or unmodeled weather dynamics.

King Abdullah University of Science and Technologytracking accuracytracking algorithmtracking methodUAV trackingUAV simulatorUAV123 dataset
IVUL-KAUST/UAVSimUE4Mueller et al (2016).pdf
  • Paper: Object Tracking Benchmark, Yi Wu et al. (2015). Its standardized single-object tracking protocols and attribute-based evaluation provide the benchmark conventions that this paper adapts to reveal how aerial conditions change tracker performance.
Cover for A Benchmark and Simulator for UAV Tracking

Abstract

In this paper, we propose a new aerial video dataset and benchmark for low altitude UAV target tracking, as well as, a photo-realistic UAV simulator that can be coupled with tracking methods. Our benchmark provides the first evaluation of many state-of-the-art and popular trackers on 123 new and fully annotated HD video sequences captured from a low-altitude aerial perspective. Among the compared trackers, we determine which ones are the most suitable for UAV tracking both in terms of tracking accuracy and run-time. The simulator can be used to evaluate tracking algorithms in real-time scenarios before they are deployed on a UAV “in the field”, as well as, generate synthetic but photo-realistic tracking datasets with automatic ground truth annotations to easily extend existing real-world datasets. Both the benchmark and simulator are made publicly available to the vision community on our website to further research in the area of object tracking from UAVs. (https://ivul.kaust.edu.sa/Pages/pub-benchmark-simulator-uav.aspx.).

Topics

tracking accuracytracking algorithmtracking methodtracking systemvideo streams

Keywords

UAV trackingUAV simulatorUAV123 datasetspatial robustness evaluationaerial tracking scenariosvisual trackinglong-term trackingtracker evaluationaerial surveillanceunmanned aerial vehicles

Table of Contents

  • 1 Introduction
  • Related Work
  • 2 Benchmark - Offline Evaluation
  • 2.1 Dataset
  • 2.2 Evaluated Algorithms
  • 2.3 Evaluation Methodology
  • 3 Simulator - Online Evaluation
  • 3.1 Setup and Limitations
  • 3.2 Novel Approaches for Evaluation
  • 3.3 Evaluation Methodology
  • 4 Experiments
  • 4.1 Benchmark Evaluation
  • 4.2 Simulator Evaluation (Quantitative and Qualitative Results)
  • 5 Conclusions and Future Work
  • References

Knowls

  1. Knowl 1 — UAV123 and UAV20L Aerial Tracking Benchmark Datasets

    experimental setup

    The UAV123 benchmark dataset is a collection of 123 high-definition aerial video sequences totaling 112,578 frames, recorded specifically for visual object tracking from a low-altitude unmanned aerial vehicle (UAV) perspective. The dataset is partitioned into three distinct subsets:

    • Set1 (103 sequences): Recorded using a professional DJI S1000 multirotor UAV flying at altitudes between 5 and 25 meters, equipped with a Panasonic GH4 camera and an Olympus M. Zuiko 12 mm f/2.0f/2.0 lens mounted on a 3-axis stabilized DJI Zenmuse Z15 gimbal. Videos were captured at frame rates between 30 and 96 FPS at resolutions from 720p to 4K, all standardized to 720p at 30 FPS. Upright bounding boxes were manually annotated at 10 FPS and linearly interpolated to 30 FPS.
    • Set2 (12 sequences): Recorded with an unstabilized board camera mounted on a low-cost UAV following other UAVs in flight. These sequences exhibit lower visual quality, lower resolution, and transmission noise.
    • Set3 (8 sequences): Synthetic sequences generated with an Unreal Engine 4 simulator from a flying UAV perspective following predefined target trajectories, annotated automatically at 30 FPS with both bounding boxes and full pixel segmentation masks.

    UAV20L Sub-dataset for Long-Term Tracking: To benchmark long-term aerial tracking, continuous uninterrupted shots that were split in the main dataset were merged, and the 20 longest sequences were selected. UAV20L contains 58,670 total frames with sequence lengths ranging from 1,717 to 5,527 frames (mean: 2,934 frames).

  2. Knowl 2 — Visual Tracking Attribute Taxonomy for Aerial Sequences

    definition

    Sequences in the UAV123 benchmark are labeled with 12 distinct binary tracking attributes to categorize specific challenges encountered in low-altitude aerial tracking:

    • Scale Variation (SV): The ratio of the target's initial bounding box area to at least one subsequent bounding box area lies outside the range [0.5,2.0][0.5, 2.0].
    • Aspect Ratio Change (ARC): The ratio of the target's aspect ratio in the first frame to that in at least one subsequent frame lies outside the range [0.5,2.0][0.5, 2.0].
    • Low Resolution (LR): At least one ground truth bounding box has an area smaller than 400 pixels (e.g., smaller than 20×2020 \times 20 pixels).
    • Fast Motion (FM): The center location of the target bounding box shifts by more than 20 pixels between two consecutive frames.
    • Camera Motion (CM): Abrupt and rapid motion of the UAV onboard camera.
    • Full Occlusion (FOC): The target object is completely occluded.
    • Partial Occlusion (POC): The target object is partially occluded.
    • Out-of-View (OV): A portion of the target extends outside the image frame.
    • Background Clutter (BC): The background adjacent to the target shares a similar visual appearance with the target.
    • Illumination Variation (IV): Significant change in the illumination incident on the target.
    • Similar Object (SOB): Objects of similar shape or category appear near the target.
    • Viewpoint Change (VC): Aerial viewpoint alteration significantly changes the target's apparent 2D appearance.

  3. Knowl 3 — Unreal Engine 4 Real-Time UAV Tracking Simulator

    model/method

    The UAV tracking simulator is built within Unreal Engine 4 (UE4) to enable real-time, closed-loop evaluation of visual tracking algorithms coupled directly with UAV aerodynamics and visual servoing:

    • Aerodynamic and Flight Modeling: The simulated UAV models the 3D geometry, weight distribution, and thrust vectors of a DJI S1000+ multirotor subjected to gravity, air resistance, and simulated wind dynamics. Proportional-Integral-Derivative (PID) controllers emulate Pixhawk flight controllers to manage flight stabilization and 3-axis gimbal visual servoing.
    • Closed-Loop Visual Servoing: External tracking algorithms implemented in C++ or MATLAB connect directly to the simulator. At each simulation step, the engine feeds the current rendered camera frame to the tracker, receives the predicted bounding box, and computes positional errors relative to the image frame center to actively update UAV position, velocity, and gimbal orientation in real time.
    • Ground Truth Generation: By utilizing UE4 post-processing custom depth buffers assigned to specific actor meshes, the simulator automatically extracts exact 2D bounding boxes, full object pixel segmentations, and 3D position/orientation trajectories for both the target and the UAV at 30 FPS.

  4. Knowl 4 — Comparison of Object Tracking Datasets

    data/table

    The table below compares the sequence counts, minimum sequence length, mean sequence length, maximum sequence length, and total frame counts between the UAV123 / UAV20L datasets and existing generic and aerial tracking benchmarks.

    Dataset UAV123 UAV20L VIVID OTB50 OTB100 TC128 VOT14 VOT15 ALOV300
    Sequences 123 20 9 51 100 129 25 60 314
    Min frames 109 1717 1301 71 71 71 171 48 19
    Mean frames 915 2934 1808 578 590 429 416 365 483
    Max frames 3085 5527 2571 3872 3872 3872 1217 1507 5975
    Total frames 112578 58670 16274 29491 59040 55346 10389 21871 151657

    UAV123 is the second largest tracking dataset by total frame count after ALOV300++, while offering significantly longer average sequence lengths (915 frames vs. 483 frames in ALOV300++ and 590 frames in OTB100). Prior aerial datasets like VIVID contain only 9 sequences and 16,274 frames, focusing on fixed-wing high-altitude vehicles. UAV20L provides 20 long sequences with a mean length of 2,934 frames, establishing a dedicated benchmark for long-term aerial drift and re-detection.

  5. Knowl 5 — UAV Tracking Evaluation Metrics and Protocols

    experimental setup

    Trackers on the UAV123 benchmark are evaluated using One-Pass Evaluation (OPE) and Spatial Robustness Evaluation (SRE):

    • Precision Metric: Measures the percentage of frames where the Euclidean distance between the predicted bounding box center and the ground truth bounding box center is within a given threshold. Trackers are ranked based on precision at a standard threshold of 20 pixels.
    • Success Metric: Measures the Intersection over Union (IoU) overlap between the predicted bounding box BtrB_{\text{tr}} and ground truth bounding box BgtB_{\text{gt}}: Overlap Score=∣Btr∩Bgt∣∣Btr∪Bgt∣\text{Overlap Score} = \frac{|B_{\text{tr}} \cap B_{\text{gt}}|}{|B_{\text{tr}} \cup B_{\text{gt}}|} Trackers are ranked using the Area Under the Curve (AUC) of the success plot, which shows the fraction of frames exceeding overlap thresholds from 0 to 1.
    • Spatial Robustness Evaluation (SRE): Measures sensitivity to initialization errors by perturbing the initial ground truth box with 4 center translations, 4 corner translations, and 4 scale modifications (80%80\%, 90%90\%, 110%110\%, and 120%120\%).
    • Workstation Environment: All 14 evaluated trackers (SRDCF, SAMF, MEEM, MUSTER, DSST, Struck, ASLA, IVT, TLD, MOSSE, CSK, OAB, KCF, and DCF) are executed with author-provided default parameters on an Intel Xeon X5675 3.07 GHz server with 48 GB RAM.

  6. Knowl 6 — Tracker Accuracy and Dominance of Scale Adaptation on UAV123

    empirical result

    Evaluating 14 trackers across the 112,578 frames of UAV123 demonstrates that aerial tracking poses challenges that significantly degrade performance compared to ground-level benchmarks (e.g., OTB100):

    • Leading Trackers: Spatially Regularized Discriminative Correlation Filter (SRDCF) achieves the highest performance in both precision (0.676 at 20 px) and success AUC (0.494), followed by SAMF, MUSTER, and DSST.
    • Scale Adaptation Requirement: Scale variation is present in 89%89\% of UAV123 sequences. Trackers with explicit scale adaptation (SRDCF, SAMF, MUSTER, DSST) consistently populate the top ranks. MEEM, which achieves the top precision on OTB100, drops significantly on UAV123 due to its lack of scale adaptation.
    • Correlation Filter Efficiency: Correlation filter-based trackers occupy the top tier in precision and success while requiring low computational cost due to Fourier-domain circular convolutions, making them suitable for onboard UAV execution.
    • Low Resolution and Camera Motion Bottlenecks: Across the benchmark, trackers suffer severe drops in performance on low-resolution sequences (such as UAV-to-UAV tracking) and sequences characterized by abrupt camera motion and aspect ratio changes.

  7. Knowl 7 — Temporal Downsampling and Tracker Latency Degradation

    empirical result

    To analyze how slow tracking speeds impact performance when intermediate frames are dropped during onboard processing, trackers were evaluated on UAV123 temporally downsampled to 10 FPS:

    • Performance Degradation: Trackers with low frame rates (such as MUSTER at 0.9 FPS, SRDCF, DSST, SAMF, and ASLA) exhibit severe score degradation. ASLA, DSST, and SAMF degrade by 21%21\% to 36%36\% in tracking score, while SRDCF, STRUCK, and MUSTER drop by 11%11\% to 15%15\%.
    • MEEM Invariance to Frame Skips: MEEM exhibits only a 7%7\% degradation in performance under 10 FPS temporal downsampling, outperforming all other trackers to rank first on UAV123@10fps. MEEM's multi-expert restoration strategy enables it to recover from large inter-frame object displacements without drifting.

  8. Knowl 8 — Long-Term Aerial Tracking Performance on UAV20L

    empirical result

    Evaluating trackers on the long sequences of UAV20L (20 sequences, mean length 2,934 frames) shows a marked drop in performance across all algorithms compared to standard UAV123 sequences, caused by accumulated model drift and occluder contamination during online model updates:

    • Top Performer: MUSTER achieves the highest performance on UAV20L with a precision score of 0.514 and success AUC of 0.359, surpassing SRDCF (precision 0.507, success AUC 0.343).
    • Long-Term Memory Mitigation: MUSTER's superior resilience on long sequences stems from its dual-store architecture: a short-term memory tracker handles local frame-to-frame motion while a long-term memory module based on keypoint matching detects tracking failures and re-initializes the target after full occlusions or severe appearance shifts.

  9. Knowl 9 — Closed-Loop Flight Dynamics and Altitude Compensation in Visual Servoing

    empirical result

    Real-time closed-loop testing of trackers (SRDCF, MEEM, SAMF, STRUCK) controlling a UAV following a moving ground vehicle (target speed up to 6 m/s, UAV speed up to 12 m/s over 3.5 minutes / ~6000 frames) reveals key dynamics between visual tracking latency and flight control:

    • Latency vs. Controller Responsiveness: Slower trackers (SRDCF and SAMF) introduce visual latency that creates lag in the estimated bounding box during rapid acceleration. When the UAV's PID flight controller is configured to be highly aggressive/responsive, tracker latency induces severe overshoot and flight instability, causing SRDCF to fail early. Conversely, faster trackers (MEEM, STRUCK) benefit from aggressive PID tuning.
    • Altitude Mitigation: Increasing UAV flight altitude widens the ground footprint per pixel, reducing apparent target motion in the image plane. This visual scaling compensates for tracker latency and significantly improves tracking precision and flight stability for slower trackers attempting to shadow fast targets.
    • Course Completion: Under default conservative PID settings, SRDCF was the only evaluated tracker capable of completing the entire 3.5-minute flight course, despite tracking drift toward salient vehicle sub-parts.

Coverage note — Omitted detailed mathematical descriptions and internal algorithmic formulations of third-party baseline trackers (e.g., OAB, IVT, CSK, KCF, TLD) as they are prior work cited by the paper rather than original contributions.

References

  1. 1.Babenko, B., Yang, M.H., Belongie, S.: Visual tracking with online multiple instance learning. IEEE Trans. Pattern Anal. Mach. Intell. 33(8), 1619–1632 (2010)
  2. 2.Battaglia, P.W., Hamrick, J.B., Tenenbaum, J.B.: Simulation as an engine of physical scene understanding. Proc. Natl. Acad. Sci. 110(45), 18327–18332 (2013). http://www.pnas.org/content/110/45/18327.abstract
  3. 3.Bibi, A., Ghanem, B.: Multi-template scale-adaptive kernelized correlation filters. In: 2015 IEEE International Conference on Computer Vision Workshop (ICCVW), pp. 613–620, December 2015
  4. 4.Bibi, A., Mueller, M., Ghanem, B.: Target response adaptation for correlation filter tracking. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9910, pp. 419–433. Springer, Switzerland (2016)
  5. 5.Bolme, D.S., Beveridge, J.R., Draper, B.A., Lui, Y.M.: Visual object tracking using adaptive correlation filters. In: 2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2544–2550, June 2010
  6. 6.Collins, R., Zhou, X., Teh, S.K.: An open source tracking testbed and evaluation web site. In: IEEE International Workshop on Performance Evaluation of Tracking and Surveillance (PETS 2005), January 2005
  7. 7.Danelljan, M., Hager, G., Shahbaz Khan, F., Felsberg, M.: Learning spatially regularized correlation filters for visual tracking. In: The IEEE International Conference on Computer Vision (ICCV), December 2015
  8. 8.Danelljan, M., Hger, G., Shahbaz Khan, F., Felsberg, M.: Accurate scale estimation for robust visual tracking. In: Proceedings of the British Machine Vision Conference. BMVA Press (2014)
  9. 9.Fu, C., Carrio, A., Olivares-Mendez, M., Suarez-Fernandez, R., Campoy, P.: Robust real-time vision-based aircraft tracking from unmanned aerial vehicles. In: 2014 IEEE International Conference on Robotics and Automation (ICRA), pp. 5441–5446, May 2014
  10. 10.Gaszczak, A., Breckon, T.P., Han, J.: Real-time people and vehicle detection from UAV imagery. In: R¨oning, J., Casasent, D.P., Hall, E.L. (eds.) IST/SPIE Electronic Imaging, vol. 7878, pp. 78780B-1–78780B-13. International Society for Optics and Photonics, January 2011
  11. 11.Grabner, H., Grabner, M., Bischof, H.: Real-time tracking via on-line boosting. In: Proceedings of the British Machine Vision Conference, pp. 6.1–6.10. BMVA Press (2006). doi:10.5244/C.20.6
  12. 12.Hamalainen, P., Eriksson, S., Tanskanen, E., Kyrki, V., Lehtinen, J.: Online motion synthesis using sequential monte carlo. ACM Trans. Graph. 33(4), 51:1–51:12. http://doi.acm.org/10.1145/2601097.2601218
  13. 13.Hare, S., Saffari, A., Torr, P.H.S.: Struck: structured output tracking with kernels. In: 2011 International Conference on Computer Vision, pp. 263–270. IEEE, November 2011
  14. 14.Hattori, H., Naresh Boddeti, V., Kitani, K.M., Kanade, T.: Learning scene-specific pedestrian detectors without real data. In: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2015
  15. 15.Hejrati, M., Ramanan, D.: Analysis by synthesis: 3D object recognition by object reconstruction. In: 2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2449–2456, June 2014
  16. 16.Henriques, J.F., Caseiro, R., Martins, P., Batista, J.: High-speed tracking with kernelized correlation filters. IEEE Trans. Pattern Anal. Mach. Intell. 37(3), 583–596 (2015)
  17. 17.Henriques, J.F., Caseiro, R., Martins, P., Batista, J.: Exploiting the circulant structure of tracking-by-detection with kernels. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012. LNCS, vol. 7575, pp. 702–715. Springer, Heidelberg (2012). doi:10.1007/978-3-642-33765-9 50
  18. 18.Hong, Z., Chen, Z., Wang, C., Mei, X., Prokhorov, D., Tao, D.: Multi-store tracker (MUSTer): a cognitive psychology inspired approach to object tracking. In: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 749–758, June 2015
  19. 19.Jia, X., Lu, H., Yang, M.H.: Visual tracking via adaptive structural local sparse appearance model. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1822–1829, June 2012
  20. 20.Ju, E., Won, J., Lee, J., Choi, B., Noh, J., Choi, M.G.: Data-driven control of flapping flight. ACM Trans. Graph. 32(5), 151:1–151:12. http://doi.acm.org/10.1145/2516971.2516976
  21. 21.Kalal, Z., Mikolajczyk, K., Matas, J.: Tracking-learning-detection. IEEE Trans. Pattern Anal. Mach. Intell. 34(7), 1409–1422 (2011)
  22. 22.Kendall, A., Salvapantula, N., Stol, K.: On-board object tracking control of a quadcopter with monocular vision. In: 2014 International Conference on Unmanned Aircraft Systems (ICUAS), pp. 404–411, May 2014
  23. 23.Kristan, M., et al.: The Visual Object Tracking VOT2014 challenge results. In: Agapito, L., Bronstein, M.M., Rother, C. (eds.) ECCV 2014. LNCS, vol. 8926, pp. 191–217. Springer, Switzerland (2015). doi:10.1007/978-3-319-16181-5 14
  24. 24.Li, A., Lin, M., Wu, Y., Yang, M.H., Yan, S.: NUS-PRO: a new visual tracking challenge. IEEE Trans. Pattern Anal. Mach. Intell. 38(2), 335–349 (2016)
  25. 25.Liang, P., Blasch, E., Ling, H.: Encoding color information for visual tracking: algorithms and benchmark. IEEE Trans. Image Process. 24(12), 5630–5644 (2015)
  26. 26.Liang, P., Blasch, E., Ling, H.: Encoding color information for visual tracking: algorithms and benchmark. IEEE Image Process. 24, 5630–5644 (2015). http://ieeexplore.ieee.org/xpls/abs all.jsp?arnumber=7277070
  27. 27.Lim, H., Sinha, S.N.: Monocular localization of a moving person onboard a quadrotor MAV. In: 2015 IEEE International Conference on Robotics and Automation (ICRA), pp. 2182–2189, May 2015
  28. 28.Mueller, M., amd Neil Smith, G.S., Ghanem, B.: Persistent aerial tracking system for UAVs. In: 2016 IEEE/RSJ International Conference Intelligent Robots and Systems (IROS), October 2016
  29. 29.Naseer, T., Sturm, J., Cremers, D.: Followme: person following and gesture recognition with a quadrocopter. In: 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 624–630, November 2013
  30. 30.Nussberger, A., Grabner, H., Van Gool, L.: Aerial object tracking from an airborne platform. In: 2014 International Conference on Unmanned Aircraft Systems (ICUAS), pp. 1284–1293, May 2014
  31. 31.Papon, J., Schoeler, M.: Semantic pose using deep networks trained on synthetic RGB-D. CoRR abs/1508.00835 (2015). http://arxiv.org/abs/1508.00835
  32. 32.Pepik, B., Stark, M., Gehler, P., Schiele, B.: Teaching 3D geometry to deformable part models. In: 2012 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3362–3369, June 2012
  33. 33.Pestana, J., Sanchez-Lopez, J., Campoy, P., Saripalli, S.: Vision based GPS-denied object tracking and following for unmanned aerial vehicles. In: 2013 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), pp. 1–6, October 2013
  34. 34.Pollard, T., Antone, M.: Detecting and tracking all moving objects in wide-area aerial video. In: 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 15–22, June 2012
  35. 35.Portmann, J., Lynen, S., Chli, M., Siegwart, R.: People detection and tracking from aerial thermal views. In: 2014 IEEE International Conference on Robotics and Automation (ICRA), pp. 1794–1800, May 2014
  36. 36.Prokaj, J., Medioni, G.: Persistent tracking for wide area aerial surveillance. In: 2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1186–1193, June 2014
  37. 37.Qadir, A., Neubert, J., Semke, W., Schultz, R.: On-board visual tracking with Unmanned Aircraft System (UAS). In: Infotech@Aerospace Conferences. American Institute of Aeronautics and Astronautics, March 2011
  38. 38.Ross, D., Lim, J., Lin, R.S., Yang, M.H.: Incremental learning for robust visual tracking. Int. J. Comput. Vis. 77(1–3), 125–141 (2008)
  39. 39.Smeulders, A.W.M., Chu, D.M., Cucchiara, R., Calderara, S., Dehghan, A., Shah, M.: Visual tracking: an experimental survey. IEEE Trans. Pattern Anal. Mach. Intell. 36(7), 1442–1468 (2014)
  40. 40.Trilaksono, B.R., Triadhitama, R., Adiprawita, W., Wibowo, A., Sreenatha, A.: Hardware in the loop simulation for visual target tracking of octorotor UAV. Aircr. Eng. Aerosp. Technol. 83(6), 407–419 (2011). http://dx.doi.org/10.1108/00022661111173289
  41. 41.Wu, Y., Lim, J., Yang, M.H.: Object tracking benchmark. IEEE Trans. Pattern Anal. Mach. Intell. 37(9), 1834–1848 (2015)
  42. 42.Wu, Y., Lim, J., Yang, M.H.: Online object tracking: a benchmark. In: 2013 IEEE Conference on Computer Vision and Pattern Recognition, pp. 2411–2418. IEEE, June 2013
  43. 43.Zhang, J., Ma, S., Sclaroff, S.: MEEM: robust tracking via multiple experts using entropy minimization. In: Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T. (eds.) ECCV 2014. LNCS, vol. 8694, pp. 188–203. Springer, Switzerland (2014). doi:10.1007/978-3-319-10599-4 13
  44. 44.Zhang, T., Bibi, A., Ghanem, B.: In defense of sparse tracking: circulant sparse tracker. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2016
  45. 45.Zhang, T., Ghanem, B., Liu, S., Xu, C., Ahuja, N.: Robust visual tracking via exclusive context modeling. IEEE Trans. Cybern. 46(1), 51–63 (2016)
  46. 46.Zhang, T., Ghanem, B., Xu, C., Ahuja, N.: Object tracking by occlusion detection via structured sparse learning. In: 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp. 1033–1040, June 2013
  47. 47.Zhang, T., Liu, S., Xu, C., Yan, S., Ghanem, B., Ahuja, N., Yang, M.H.: Structural sparse tracking. In: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 150–158, June 2015

Citation

MLA
Mueller, M., et al. “A Benchmark and Simulator for UAV Tracking”. Lecture Notes in Computer Science, Springer International Publishing, 2016, pp. 445–61, https://doi.org/10.1007/978-3-319-46448-0_27.
APA
Mueller, M., Smith, N., & Ghanem, B. (2016). A Benchmark and Simulator for UAV Tracking. In Lecture Notes in Computer Science (pp. 445–461). Springer International Publishing. https://doi.org/10.1007/978-3-319-46448-0_27
Chicago
Mueller, M., N. Smith, and B. Ghanem. 2016. “A Benchmark and Simulator for UAV Tracking”. In Lecture Notes in Computer Science. Springer International Publishing. https://doi.org/10.1007/978-3-319-46448-0_27.
Harvard
Mueller, M., Smith, N. and Ghanem, B. (2016) “A Benchmark and Simulator for UAV Tracking”, Lecture Notes in Computer Science. Springer International Publishing, pp. 445–461. Available at: https://doi.org/10.1007/978-3-319-46448-0_27.
Vancouver
1. Mueller M, Smith N, Ghanem B (2016) A Benchmark and Simulator for UAV Tracking. In: Lecture Notes in Computer Science. Springer International Publishing, pp 445–461

BibTeX

@inbook{Mueller_2016, title={A Benchmark and Simulator for UAV Tracking}, ISBN={9783319464480}, ISSN={1611-3349}, url={http://dx.doi.org/10.1007/978-3-319-46448-0_27}, DOI={10.1007/978-3-319-46448-0_27}, booktitle={Computer Vision – ECCV 2016}, publisher={Springer International Publishing}, author={Mueller, Matthias and Smith, Neil and Ghanem, Bernard}, year={2016}, pages={445–461} }
Metadata:Crossref

Source Code

This paper has an official code repository available. Click below to access the source code.

View Repository

Access the Paper

This paper is available from its original source. Click below to access the PDF.

Open PDF