Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Efficient and Accurate Downfacing Visual Inertial Odometry

T0 review · 3 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A low-power, downfacing VIO pipeline using quantized ORB and a rigid-body motion model reduces indoor trajectory error by up to 3.65x over the original PX4FLOW baseline.

desk verdict Real engineering value with a headline accuracy claim that overstates what the controlled data show — fixable with a matched baseline. read the letter →

arxiv 2509.10021 v1 pith:26AJMTZU submitted 2025-09-12 cs.CV cs.ROeess.IV

Jonas Kühne , Christian Vogt , Michele Magno , Luca Benini This is my paper · ORCID
classification cs.CVcs.ROeess.IV
keywords downfacingVIOvisualinertialodometryultra-low-powerSoCRISC-VPX4FLOWORBSuperPointrigid-bodymotionestimation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that accurate visual-inertial odometry can run on ultra-low-power RISC-V SoCs, and that the accuracy comes from a simple change: replace the averaged optical-flow estimate of the classic PX4FLOW pipeline with a rigid-body motion decomposition (translation in x, y, and yaw rotation) plus a two-stage outlier rejection, then fuse with IMU data in an extended Kalman filter. The optimized pipeline, evaluated with ORB, PX4FLOW, and SuperPoint feature trackers on the GAP9 SoC, reports an average RMSE reduction of up to 3.65x over the baseline on indoor sequences. The paper also characterizes the trade-off: PX4FLOW remains competitive below 24 pixels/frame of motion at lower computational cost, while ORB handles larger displacements. If right, this narrows the gap between high-accuracy VIO that usually needs a powerful processor and lightweight implementations that fit microcontrollers, enabling onboard navigation for micro- and nano-drones.

What carries the argument

The central object is the rigid-body motion model for downfacing cameras: each feature's displacement is assumed to come from a planar translation (Δu, Δv) plus a yaw rotation (Δψ) around the image center. The pipeline estimates these parameters from tracked pixel coordinates using a two-stage outlier rejection (a histogram threshold of five pixels, then a 1.5-pixel inlier check) and solves the resulting system by singular-value decomposition of a 2x2 matrix. This motion estimate is fused with IMU acceleration and gyroscope states in an extended Kalman filter; the feature trackers (ORB with integer arithmetic, SuperPoint with 8-bit quantization, and fixed-point PX4FLOW) supply the raw flow.

What would settle it

Run the ORB-based pipeline on the same indoor sequences at 300 FPS (matching the original PX4FLOW frame rate) with the same EKF fusion; if its RMSE is no better than the baseline's, the claimed 3.65x gain is explained by frame rate, not by the rigid-body motion model. Alternatively, replace the rigid-body decomposition in the proposed pipeline with averaged flow while keeping tracker, frame rate, and filter identical; if RMSE does not worsen, the motion model is not the driver.

Watch

Extended reading notes

Core claim

The central claim is that a downfacing VIO pipeline can achieve substantially better accuracy than the original PX4FLOW implementation without leaving the microcontroller power envelope. The key to the improvement is modeling the feature displacement as a rigid-body motion: the tracker outputs per-feature optical flow, and the pipeline solves for the translation (Δu, Δv) and yaw rotation (Δψ) of the camera in a single least-squares step, after a coarse histogram-based outlier rejection and a second iterative inlier refinement. This replaces PX4FLOW's weighted average of flows and, crucially, feeds the yaw estimate into an EKF together with full IMU acceleration and gyroscope data, whereas th

Load-bearing premise

The central claim collapses if the accuracy improvement is mostly an artifact of frame rate and retuning rather than the rigid-body motion model, since the reported comparison is not matched on those variables.

Editorial extensions

If this is right

  • Indoor flights with a downfacing camera can use a sub-68 mW SoC for onboard VIO with better accuracy than the original PX4FLOW, removing the need for a ground station or high-power processor.
  • For movement speeds below 24 pixels/frame, the PX4FLOW tracker integrated into the new pipeline is a lower-cost alternative to ORB with roughly comparable accuracy.
  • For larger per-frame displacements, ORB's runtime is independent of displacement, making it the recommended choice over PX4FLOW, whose cost grows quadratically with trackable range.
  • The rigid-body motion model's benefit shows up mainly in turns: the paper reports the original PX4FLOW remains better in pure translations, but the new pipeline is more robust in square and random trajectories.
  • The integer-only ORB implementation is hardware-agnostic, so the accuracy gains can be ported to other MCU-class platforms beyond GAP9.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The 3.65x headline number should not be read as the isolated effect of the rigid-body model: the comparison baseline runs at 300 FPS with gyro-only fusion, while the new pipeline runs at 100 FPS with full IMU fusion. The paper's own ablation shows frame rate alone can swing RMSE by more than an order of magnitude, so a matched baseline would likely show a smaller, though still positive, gain.
  • A direct test of the motion model's value would be to run the same ORB tracker at the same frame rate and EKF, and toggle only the rigid-body decomposition on and off; the paper does not report this ablation for ORB.
  • The proposed approach depends on the planar-motion assumption, so it transfers most naturally to other downfacing or ground-relative platforms (delivery robots, AR glasses with downward cameras) rather than to general six-degree-of-freedom VIO.
  • The PX4FLOW/ORB crossover at roughly 24 pixels/frame suggests an adaptive scheduler could switch trackers based on estimated optical flow, saving power during slow segments and switching to ORB during fast maneuvers.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript presents a downfacing visual-inertial odometry (VIO) pipeline targeting ultra-low-power RISC-V SoCs, with three interchangeable feature trackers (ORB, SuperPoint, and a parallelized PX4FLOW derivative) feeding a common template that performs rigid-body motion estimation with outlier rejection and EKF-based IMU fusion. The system is implemented and profiled on the GAP9 SoC, and accuracy is evaluated on indoor sequences with Vicon ground truth and outdoor sequences with GPS-RTK ground truth, reporting RMSE and relative translation error. An ablation studies PX4FLOW frame rate and search range. The headline claim is that the optimized pipeline achieves an average RMSE reduction of up to 3.65x over the original PX4FLOW baseline when using the ORB tracker.

Significance. If the accuracy claim holds, this is a valuable systems contribution showing that feature-based VIO with competitive accuracy can run on sub-68 mW RISC-V platforms. The paper's strengths are concrete: the implementation is open-sourced, real-world benchmarks with two ground-truth systems are reported with standard deviations, and a focused ablation of PX4FLOW frame rate and search range is included. The central weakness is experimental attribution: the headline 3.65x comparison varies frame rate, IMU fusion, and motion model simultaneously, and the paper's own ablation data show frame-rate effects large enough to dominate the reported factor. The work is empirical rather than derivational, so this attribution issue is load-bearing for the main claim.

major comments (3)
  1. [Abstract; Table III; Table VI; §V-C] The 3.65x RMSE-reduction claim compares pipelines that differ in three independent respects: feature tracker (ORB vs PX4FLOW), frame rate (100 vs 300 FPS), and estimator (rigid-body + full IMU EKF vs averaged flow + gyro-only). Table VI shows that changing only the frame rate from 100 to 300 FPS while keeping the PX4FLOW tracker and the template pipeline increases RMSE by up to an order of magnitude (e.g., sequence 05: 0.320 m to 3.582 m), and §V-C concedes the outlier thresholds and Kalman gains would need retuning at 300 FPS. Thus the 3.65x factor cannot be attributed to the proposed rigid-body motion model or the ORB tracker without a matched-baseline ablation, e.g., the original PX4FLOW estimator at 100 FPS, or the proposed template with gyro-only IMU, or properly retuned 300 FPS PX4FLOW.
  2. [§IV-D; Abstract] The accuracy results are reported as run 'at a hypothetical 100 FPS on the GVSoC' (§IV-D), while the abstract and introduction state that the pipeline was 'implemented on the SoC for real-world validation.' Please clarify whether the RMSE numbers in Tables III–VI were produced by actual GAP9 execution or by simulation. If they are from GVSoC, the SoC-validation claim should be limited to profiling and real-time capability, with accuracy demonstrated on the simulator.
  3. [Table III] The 'when using the ORB feature tracker' qualifier is also undercut by Table III: the modified PX4FLOW variant in the same template at the same 100 FPS achieves a lower average RMSE (approximately 0.681 m) than ORB (approximately 0.752 m). This indicates that the improvement over the original PX4FLOW baseline is not primarily due to the choice of ORB. The paper should state explicitly which component (rigid-body model, IMU fusion, or tracker) drives the reported gain, or rephrase the claim to refer only to the full pipeline versus the baseline.
minor comments (5)
  1. [§II-A] Typo: 'Extended Klaman Filters' should be 'Extended Kalman Filters'.
  2. [§V-A2] Typo: 'locally optmized' should be 'locally optimized'; also 'kcyles' should be 'kcycles'.
  3. [§V-A1] The text states the Kalman Filtering module requires '156 cycles', but Table II reports 156.0 kCycles. Please correct the unit.
  4. [Table III] Header uses 'Framerate' and nonstandard spacing in 'Max Movement ±32pixel'; consider uniform formatting and units.
  5. [§I] Minor spacing issues in 'UA V' and 'PIX4FLOW' (should be 'UAV' and 'PX4FLOW') appear in a few places; a copyedit pass is recommended.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: the accuracy claim is an empirical measurement against external ground truth, and the self-citations are baseline code sources rather than load-bearing premises.

full rationale

The paper makes no first-principles derivation that could be circular: VIO accuracy is measured on recorded indoor/outdoor sequences (Vicon and GPS-RTK ground truth) with the standard RPG trajectory evaluation toolkit [33], and the 3.65x RMSE factor is the ratio of two empirical averages computed from Table III (original PX4FLOW average RMSE 2.74 m vs ORB-in-template average RMSE 0.75 m). No fitted parameter is renamed as a prediction: the Kalman gains, outlier thresholds, and frame rates are configuration choices, and the paper explicitly concedes in Section V-C that retuning would be needed at 300 FPS, which is a comparison-validity issue rather than a circular-reasoning issue. The template pipeline's rigid-body motion decomposition is a standard textbook model [17],[18], not an ansatz smuggled in by self-citation. The only self-citations ([9], [14]) identify the parallelized PX4FLOW baseline implementation and its locally optimized derivative; these are benchmark artifacts, and the baseline is also the independent original PX4FLOW design [8]. The paper does not invoke a self-authored uniqueness theorem, nor does any central claim reduce by construction to its own inputs. The comparison is confounded (frame rate, IMU state, motion model) and the paper's own ablation shows frame-rate sensitivity, but that is a threat to internal validity, not circularity. Therefore no circular step is present; score reflects only a minor, non-load-bearing self-citation in the baseline choice.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The quantitative claims rest on several hand-tuned parameters (outlier thresholds, Kalman gains, matching thresholds) and on the planar-motion domain assumption. No new physical entities are introduced. The free parameters are not fitted to the benchmark dataset in a scientifically circular way, but they are undisclosed tuning choices that affect the headline results.

free parameters (5)
  • Outlier rejection thresholds = 5 px (first pass), 1.5 px (second pass)
    Hand-chosen in Section III-A3; no justification from data or theory, and the paper states the pipeline would need retuning for different frame rates.
  • Kalman filter gains = not reported
    Section V-C admits the template pipeline needs tuning of gains when frame rate changes; gains are not disclosed and directly affect RMSE.
  • ORB matching Hamming distance threshold = 20
    Chosen in Section III-B5; determines which feature pairs are treated as matches and influences the flow estimates.
  • FAST/Harris detection thresholds with hysteresis = target 150-200 descriptors, upper bound 512
    Section V-A2; thresholds are adapted per texture to keep compute roughly constant, affecting feature selection and tracking quality.
  • Harris M quantization scale factor = 2^-11
    Section III-B2; scaling choice for integer implementation, stated not to affect accuracy for thresholds well above zero.
assumptions (6)
  • domain assumption The camera is downfacing and the motion is planar (translation parallel to ground plus yaw rotation)
    Invoked in Section II-B and III-A; reduces the estimation problem from 6-DoF to 4-DoF. If roll, pitch, or height changes are significant, the rigid-body decomposition is invalid.
  • standard math The rigid-body motion model from [17],[18] describes the observed optical flow
    Used in Section III-A3; standard planar motion model, assumed to hold for the tested scenes.
  • standard math The Extended Kalman Filter fuses visual and IMU states without significant model error
    Section III-A4 relies on standard EKF behavior [28]; assumes noise characteristics and state propagation are appropriate.
  • domain assumption Vicon motion capture and GPS-RTK provide accurate ground truth
    Section IV-B; RMSE is computed against these references, but no ground truth error analysis is provided.
  • domain assumption The pretrained SuperPoint checkpoint generalizes to downfacing drone imagery
    Section III-C uses the Magic Leap checkpoint without fine-tuning; quantization further degrades it (mean localization error 1.71 px, cosine similarity 0.91).
  • domain assumption The ToF sensor provides a valid reference height measurement
    Section III-A1; assumes the floor is the dominant reflective surface and that the accuracy setting (±0.15%) is adequate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient and Accurate Downfacing Visual Inertial Odometry." pith.science (2026). https://pith.science/paper/26AJMTZU

@misc{pith2026250910021,
  author       = {Pith},
  title        = {Pith review of: Efficient and Accurate Downfacing Visual Inertial Odometry},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/26AJMTZU}},
  note         = {Machine review of arXiv:2509.10021}
}
read the original abstract

Visual Inertial Odometry (VIO) is a widely used computer vision method that determines an agent's movement through a camera and an IMU sensor. This paper presents an efficient and accurate VIO pipeline optimized for applications on micro- and nano-UAVs. The proposed design incorporates state-of-the-art feature detection and tracking methods (SuperPoint, PX4FLOW, ORB), all optimized and quantized for emerging RISC-V-based ultra-low-power parallel systems on chips (SoCs). Furthermore, by employing a rigid body motion model, the pipeline reduces estimation errors and achieves improved accuracy in planar motion scenarios. The pipeline's suitability for real-time VIO is assessed on an ultra-low-power SoC in terms of compute requirements and tracking accuracy after quantization. The pipeline, including the three feature tracking methods, was implemented on the SoC for real-world validation. This design bridges the gap between high-accuracy VIO pipelines that are traditionally run on computationally powerful systems and lightweight implementations suitable for microcontrollers. The optimized pipeline on the GAP9 low-power SoC demonstrates an average reduction in RMSE of up to a factor of 3.65x over the baseline pipeline when using the ORB feature tracker. The analysis of the computational complexity of the feature trackers further shows that PX4FLOW achieves on-par tracking accuracy with ORB at a lower runtime for movement speeds below 24 pixels/frame.

Figures

Figures reproduced from arXiv: 2509.10021 by the authors.

Figure 1
Figure 1. In this work, we present a downfacing VIO pipeline that is [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Coordinate systems of the VO and IMU relative to the data collection [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 1
Figure 1. 1) Sensor Readout: The three sensors (camera, IMU, and ToF sensor) are read out by the fabric controller core. The camera and IMU are read out at the same rate, which depends on the processing time required by the selected feature tracker. The ToF sensor, which mainly provides reference height measurements, is configured in its max range mode, allowing a read-out rate of 6.94 Hz. Once the sensor data is available, i… view at source ↗
Figures from the paper (4 more)
Figure 3
Figure 3. Figure 3: Physical setup for data collection, with the camera, IMU, and ToF [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 4
Figure 4. Figure 4: Data collection setup overview with different floor types and motion [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Real world example trajectory (Sand) including the ground truth, as [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The runtime complexities of the template pipeline in combination with [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 1 linked inside Pith

  1. [1]

    Sg-vio: Monocular visual-inertial odometry with tightly coupled struc- tural lines and gravity to avoid degeneracy,

    H. Yao, Y . Ma, P. Li, C. Zhai, J. Song, M. Ouyang, Z. Dai, and X. Zhu, “Sg-vio: Monocular visual-inertial odometry with tightly coupled struc- tural lines and gravity to avoid degeneracy,”IEEE Internet of Things Journal, 2024

    2024

  2. [2]

    A deep learning- based monocular vo/pdr integrated indoor localization algorithm using smartphone,

    Y . Lin, K. Yu, F. Zhu, M. Chao, and J. Dong, “A deep learning- based monocular vo/pdr integrated indoor localization algorithm using smartphone,”IEEE Internet of Things Journal, 2024

    2024

  3. [3]

    Llio: Lightweight learned inertial odometer,

    Y . Wang, J. Kuang, X. Niu, and J. Liu, “Llio: Lightweight learned inertial odometer,”IEEE Internet of Things Journal, vol. 10, no. 3, pp. 2508– 2518, 2022

    2022

  4. [4]

    A comprehensive survey of visual slam algorithms,

    A. Macario Barros, M. Michel, Y . Moline, G. Corre, and F. Carrel, “A comprehensive survey of visual slam algorithms,”Robotics, vol. 11, no. 1, p. 24, 2022

    2022

  5. [5]

    Gap9shield: A 150gops ai-capable ultra-low power module for vision and ranging applications on nano- drones,

    H. M ¨uller, V . Kartsch, and L. Benini, “Gap9shield: A 150gops ai-capable ultra-low power module for vision and ranging applications on nano- drones,”arXiv preprint arXiv:2407.13706, 2024

    arXiv 2024

  6. [6]

    Orb: An efficient alternative to sift or surf,

    E. Rublee, V . Rabaud, K. Konolige, and G. Bradski, “Orb: An efficient alternative to sift or surf,” in2011 International conference on computer vision. Ieee, 2011, pp. 2564–2571

    2011

  7. [7]

    Superpoint: Self- supervised interest point detection and description,

    D. DeTone, T. Malisiewicz, and A. Rabinovich, “Superpoint: Self- supervised interest point detection and description,” inProceedings of the IEEE conference on computer vision and pattern recognition workshops, 2018, pp. 224–236

    2018

  8. [8]

    An open source and open hardware embedded metric optical flow cmos camera for in- door and outdoor applications,

    D. Honegger, L. Meier, P. Tanskanen, and M. Pollefeys, “An open source and open hardware embedded metric optical flow cmos camera for in- door and outdoor applications,” in2013 IEEE International Conference on Robotics and Automation. IEEE, 2013, pp. 1736–1741

    2013

Show all 36 references
  1. [9]

    Parallelizing optical flow estima- tion on an ultra-low power risc-v cluster for nano-uav navigation,

    J. K ¨uhne, M. Magno, and L. Benini, “Parallelizing optical flow estima- tion on an ultra-low power risc-v cluster for nano-uav navigation,” in 2022 IEEE International Symposium on Circuits and Systems (ISCAS). IEEE, 2022, pp. 301–305

    2022

  2. [10]

    Autonomous drone racing: A survey,

    D. Hanover, A. Loquercio, L. Bauersfeld, A. Romero, R. Penicka, Y . Song, G. Cioffi, E. Kaufmann, and D. Scaramuzza, “Autonomous drone racing: A survey,”IEEE Transactions on Robotics, 2024

    2024

  3. [11]

    Stereo visual inertial odometry for robots with limited computational resources,

    S. Bahnam, S. Pfeiffer, and G. C. de Croon, “Stereo visual inertial odometry for robots with limited computational resources,” in2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2021, pp. 9154–9159

    2021

  4. [12]

    Orb- neuroslam: A brain-inspired 3d slam system based on orb features,

    D. Shen, G. Liu, T. Li, F. Yu, F. Gu, K. Xiao, and X. Zhu, “Orb- neuroslam: A brain-inspired 3d slam system based on orb features,” IEEE Internet of Things Journal, 2023

    2023

  5. [13]

    Orb-slam3: An accurate open-source library for visual, visual– inertial, and multimap slam,

    C. Campos, R. Elvira, J. J. G. Rodr ´ıguez, J. M. Montiel, and J. D. Tard´os, “Orb-slam3: An accurate open-source library for visual, visual– inertial, and multimap slam,”IEEE Transactions on Robotics, vol. 37, no. 6, pp. 1874–1890, 2021

    2021

  6. [14]

    Low latency visual inertial odom- etry with on-sensor accelerated optical flow for resource-constrained uavs,

    J. K ¨uhne, M. Magno, and L. Benini, “Low latency visual inertial odom- etry with on-sensor accelerated optical flow for resource-constrained uavs,”IEEE Sensors Journal, 2024

    2024

  7. [15]

    Machine learning for microcontroller-class hardware: A review,

    S. S. Saha, S. S. Sandha, and M. Srivastava, “Machine learning for microcontroller-class hardware: A review,”IEEE Sensors Journal, vol. 22, no. 22, pp. 21 362–21 390, 2022

    2022

  8. [16]

    Picovo: A lightweight rgb- d visual odometry targeting resource-constrained iot devices,

    Y . He, Y . Wang, C. Liu, and L. Zhang, “Picovo: A lightweight rgb- d visual odometry targeting resource-constrained iot devices,” in2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 5567–5573

    2021

  9. [17]

    R. A. Tenenbaum,Fundamentals of applied dynamics. Springer Science & Business Media, 2006

    2006

  10. [18]

    Hierarchical model-based motion estimation,

    J. R. Bergen, P. Anandan, K. J. Hanna, and R. Hingorani, “Hierarchical model-based motion estimation,” inComputer Vision—ECCV’92: Sec- ond European Conference on Computer Vision Santa Margherita Ligure, Italy, May 19–22, 1992 Proceedings 2. Springer, 1992, pp. 237–252

    1992

  11. [19]

    Direct sparse visual- inertial odometry using dynamic marginalization,

    L. V on Stumberg, V . Usenko, and D. Cremers, “Direct sparse visual- inertial odometry using dynamic marginalization,” in2018 IEEE Inter- national Conference on Robotics and Automation (ICRA). IEEE, 2018, pp. 2510–2517

    2018

  12. [20]

    A multi-state constraint kalman filter for vision-aided inertial navigation,

    A. I. Mourikis and S. I. Roumeliotis, “A multi-state constraint kalman filter for vision-aided inertial navigation,” inProceedings 2007 IEEE international conference on robotics and automation. IEEE, 2007, pp. 3565–3572

    2007

  13. [21]

    Keyframe-based visual–inertial odometry using nonlinear optimiza- tion,

    S. Leutenegger, S. Lynen, M. Bosse, R. Siegwart, and P. Furgale, “Keyframe-based visual–inertial odometry using nonlinear optimiza- tion,”The International Journal of Robotics Research, vol. 34, no. 3, pp. 314–334, 2014

    2014

  14. [22]

    Vins-mono: A robust and versatile monocular visual-inertial state estimator,

    T. Qin, P. Li, and S. Shen, “Vins-mono: A robust and versatile monocular visual-inertial state estimator,”IEEE transactions on robotics, vol. 34, no. 4, pp. 1004–1020, 2018

    2018

  15. [23]

    Robust visual inertial odometry using a direct ekf-based approach,

    M. Bloesch, S. Omari, M. Hutter, and R. Siegwart, “Robust visual inertial odometry using a direct ekf-based approach,” in2015 IEEE/RSJ international conference on intelligent robots and systems (IROS). IEEE, 2015, pp. 298–304

    2015

  16. [24]

    Multi- aperture visual velocity measurement method based on biomimetic compound-eye for uavs,

    C. Shen, X. Zhao, X. Wu, H. Cao, C. Wang, J. Tang, and J. Liu, “Multi- aperture visual velocity measurement method based on biomimetic compound-eye for uavs,”IEEE Internet of Things Journal, 2023

    2023

  17. [25]

    Cmos image sensor technology advances for mobile devices,

    R. J. Gove, “Cmos image sensor technology advances for mobile devices,” inHigh Performance Silicon Imaging. Elsevier, 2020, pp. 185–240

    2020

  18. [26]

    Robust stereo visual inertial odometry for fast autonomous flight,

    K. Sun, K. Mohta, B. Pfrommer, M. Watterson, S. Liu, Y . Mulgaonkar, C. J. Taylor, and V . Kumar, “Robust stereo visual inertial odometry for fast autonomous flight,”IEEE Robotics and Automation Letters, vol. 3, no. 2, pp. 965–972, 2018

    2018

  19. [27]

    A heterogeneous risc-v based soc for secure nano-uav navigation,

    L. Valente, A. Nadalini, A. H. C. Veeran, M. Sinigaglia, B. S ´a, N. Wistoff, Y . Tortorella, S. Benatti, R. Psiakis, A. Kulmalaet al., “A heterogeneous risc-v based soc for secure nano-uav navigation,”IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 71, no. 5...

    2024

  20. [28]

    Kalman and extended kalman filters: Concept, derivation and properties,

    M. I. Ribeiro, “Kalman and extended kalman filters: Concept, derivation and properties,”Institute for Systems and Robotics, vol. 43, no. 46, pp. 3736–3741, 2004

    2004

  21. [29]

    Machine learning for high-speed corner detection,

    E. Rosten and T. Drummond, “Machine learning for high-speed corner detection,” inComputer Vision–ECCV 2006: 9th European Conference on Computer Vision, Graz, Austria, May 7-13, 2006. Proceedings, Part I 9. Springer, 2006, pp. 430–443

    2006

  22. [30]

    A combined corner and edge detector,

    C. Harris, M. Stephenset al., “A combined corner and edge detector,” in Alvey vision conference, vol. 15, no. 50. Citeseer, 1988, pp. 10–5244

    1988

  23. [31]

    An analysis and implementation of the harris corner detector,

    J. S ´anchez, N. Monz ´on, and A. Salgado De La Nuez, “An analysis and implementation of the harris corner detector,”Image Processing On Line, 2018

    2018

  24. [32]

    Canton of zurich, rgb/infrared geodata 2021/22,

    “Canton of zurich, rgb/infrared geodata 2021/22,” https://opendata.swiss/ de/dataset/orthofoto-fruhjahr-rgb-infrarot-2021-22, accessed: 2025-01- 20

    2021

  25. [33]

    A tutorial on quantitative trajectory eval- uation for visual (-inertial) odometry,

    Z. Zhang and D. Scaramuzza, “A tutorial on quantitative trajectory eval- uation for visual (-inertial) odometry,” in2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018, pp. 7244–7251

    2018

  26. [34]

    Qr-pulp: Streamlining qr decomposition for risc-v parallel ultra-low-power platforms,

    A. Kiamarzi, D. Rossi, and G. Tagliavini, “Qr-pulp: Streamlining qr decomposition for risc-v parallel ultra-low-power platforms,” inPro- ceedings of the 21st ACM International Conference on Computing Frontiers, 2024, pp. 147–154

    2024

  27. [35]

    On-device self- supervised learning of visual perception tasks aboard hardware-limited nano-quadrotors,

    E. Cereda, M. Rusci, A. Giusti, and D. Palossi, “On-device self- supervised learning of visual perception tasks aboard hardware-limited nano-quadrotors,” in2024 IEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 10 118–10 124

    2024

  28. [36]

    Navion: A 2-mw fully integrated real-time visual-inertial odometry accelerator for autonomous navigation of nano drones,

    A. Suleiman, Z. Zhang, L. Carlone, S. Karaman, and V . Sze, “Navion: A 2-mw fully integrated real-time visual-inertial odometry accelerator for autonomous navigation of nano drones,”IEEE Journal of Solid-State Circuits, vol. 54, no. 4, pp. 1106–1119, 2019. 12 Jonas K ¨uhne(Gra...

    2019

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.