ABSTRACT
The asynchronous event stream of event camera output overcomes exposure problems caused by dramatic changes in ambient light and motion blur caused by high-speed motion, which are common challenges with traditional cameras. With the increasing number of events per second delivered by event cameras, a faster feature extraction method is necessary to process large amounts of events to take advantage of event cameras for various computer vision tasks. We propose UCED-Detector, an event frame-based corner event detector that can detect features in event streams at three times the speed of the SOTA method. Firstly, we use events captured in the past to remove noise events from the current event stream. The events in the circular mask around the event to be detected are then constructed as event pairs and mark the events whose timestamps are one threshold larger than the other event in the event pair. Finally, the marked adjacent events are connected into arcs, and whether the event to be detected is a feature corner event is judged according to the arc length. To evaluate the performance of our proposed approach, we conducted extensive experiments on datasets of event cameras. The results show that our method reduces the detection time to one-third of the SOTA method and reduces the average processing time per event from 0.15 milliseconds to 0.04 milliseconds.
- Campos C, Elvira R, Rodríguez J J G, Orb-slam3: An accurate open-source library for visual, visual–nertial, and multimap slam [J]. IEEE Transactions on Robotics, 2021, 37(6): 1874-1890.Google ScholarCross Ref
- Qin T, Li P, Shen S. Vins-mono: A robust and versatile monocular visual-inertial state estimator [J]. IEEE Transactions on Robotics, 2018, 34(4): 1004-1020.Google ScholarDigital Library
- Bescos B, Fácil J M, Civera J, DynaSLAM: Tracking, mapping, and inpainting in dynamic scenes [J]. IEEE Robotics and Automation Letters, 2018, 3(4): 4076-4083.Google ScholarCross Ref
- Yang S, Scherer S. Cubeslam: Monocular 3-d object slam[J]. IEEE Transactions on Robotics, 2019, 35(4): 925-938.Google ScholarDigital Library
- Mur-Artal R, Tardós J D. Orb-slam2: An open-source slam system for monocular, stereo, and rgb-d cameras [J]. IEEE transactions on robotics, 2017, 33(5): 1255-1262.Google Scholar
- Zhu Z, Peng S, Larsson V, Nice-slam: Neural implicit scalable encoding for slam[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022, 12786-12796.Google Scholar
- Feng S, Sun H, Yan X, Dense reinforcement learning for safety validation of autonomous vehicles [J]. Nature, 2023, 615(7953): 620-627.Google ScholarCross Ref
- Serrano-Gotarredona T, Linares-Barranco B. A 128$\,\times $128 1.5\% Contrast Sensitivity 0.9\% FPN 3 μs Latency 4 mW Asynchronous Frame-Free Dynamic Vision Sensor Using Transimpedance Preamplifiers [J]. IEEE Journal of Solid-State Circuits, 2013, 48(3): 827-838.Google ScholarCross Ref
- Brandli C, Berner R, Yang M, A 240× 180 130 db 3 μs latency global shutter spatiotemporal vision sensor [J]. IEEE Journal of Solid-State Circuits, 2014, 49(10): 2333-2341.Google ScholarCross Ref
- Huang K, Zhang S, Zhang J, Event-based Simultaneous Localization and Mapping: A Comprehensive Survey [J]. arXiv preprint arXiv:2304.09793, 2023.Google Scholar
- Gallego G, Delbrück T, Orchard G, Event-based vision: A survey [J]. IEEE transactions on pattern analysis and machine intelligence, 2020, 44(1): 154-180.Google Scholar
- Chen P, Guan W, Lu P. Esvio: Event-based stereo visual inertial odometry [J]. IEEE Robotics and Automation Letters, 2023.Google Scholar
- Zuo Y F, Yang J, Chen J, Devo: Depth-event camera visual odometry in challenging conditions[C]//2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022: 2179-2185.Google Scholar
- Clady X, Ieng S H, Benosman R. Asynchronous event-based corner detection and matching [J]. Neural Networks, 2015, 66: 91-106.Google ScholarDigital Library
- Moeys D P, Li C, Martel J N P, Color temporal contrast sensitivity in dynamic vision sensors[C]//2017 IEEE International Symposium on Circuits and Systems (ISCAS). IEEE, 2017, 1-4.Google Scholar
- Huang J, Guo M, Chen S. A dynamic vision sensor with direct logarithmic output and full-frame picture-on-demand[C]//2017 IEEE International Symposium on Circuits and Systems (ISCAS). IEEE, 2017, 1-4.Google Scholar
- Huang J, Guo M, Wang S, A motion sensor with on-chip pixel rendering module for optical flow gradient extraction[C]//2018 IEEE International Symposium on Circuits and Systems (ISCAS). IEEE, 2018, 1-5.Google Scholar
- Benosman R, Ieng S H, Clercq C, Asynchronous frameless event-based optical flow [J]. Neural Networks, 2012, 27: 32-37.Google ScholarDigital Library
- Alzugaray I, Chli M. ACE: An efficient asynchronous corner tracker for event cameras[C]//2018 International Conference on 3D Vision (3DV). IEEE, 2018, 653-661.Google Scholar
- Manderscheid J, Sironi A, Bourdis N, Speed invariant time surface for learning to detect corner points with event-based cameras[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2019, 10245-10254.Google Scholar
- Alzugaray I, Chli M. Asynchronous corner detection and tracking for event cameras in real time[J]. IEEE Robotics and Automation Letters, 2018, 3(4): 3177-3184.Google ScholarCross Ref
- Brändli C, Strubel J, Keller S, ELiSeD–An event-based line segment detector[C]//2016 Second International Conference on Event-based Control, Communication, and Signal Processing (EBCCSP). IEEE, 2016, 1-7.Google Scholar
- Glover A, Bartolozzi C. Event-driven ball detection and gaze fixation in clutter[C]//2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2016, 2203-2208.Google Scholar
- Vasco V, Glover A, Bartolozzi C. Fast event-based Harris corner detection exploiting the advantages of event-driven cameras[C]//2016 IEEE/RSJ international conference on intelligent robots and systems (IROS). IEEE, 2016, 4144-4149.Google Scholar
- Mueggler E, Bartolozzi C, Scaramuzza D. Fast event-based corner detection [J]. 2017.Google ScholarCross Ref
- Li R, Shi D, Zhang Y, Fa-harris: A fast and asynchronous corner detector for event cameras[C]//2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2019, 6223-6229.Google Scholar
- Zhao J, Su L, Wang X, DTFS-eHarris: A High Accuracy Asynchronous Corner Detector for Event Cameras in Complex Scenes [J]. Applied Sciences, 2023, 13(9): 5761.Google ScholarCross Ref
- Mueggler E, Rebecq H, Gallego G, The event-camera dataset and simulator: Event-based data for pose estimation, visual odometry, and SLAM[J]. The International Journal of Robotics Research, 2017, 36(2): 142-149.Google ScholarDigital Library
Index Terms
- UCED-Detector: An Ultra-fast Corner Event Detector for Event Camera in Complex Scenes: UCED-Detector: Ultra-fast Detector An ultra-fast detector for detecting feature corner events in a high-speed event stream
Recommendations
Stream reasoning and complex event processing in ETALIS
On linked spatiotemporal data and geo-ontologiesAddressing dynamics and notifications in the Semantic Web realm has recently become an important area of research. Run time data is continuously generated by multiple social networks, sensor networks, various on-line services and so forth. How to get ...
Digital Video Event Detector Framework for Surveillance Applications
AVSS '09: Proceedings of the 2009 Sixth IEEE International Conference on Advanced Video and Signal Based SurveillanceThe paper introduces a video surveillance and event detection framework and application for semi-supervised surveillance use. The system’s intended use is in automatic mode on camera feeds that are not actively watched by surveillance personnel, and ...
Stream Event Detection: A Unified Framework for Mining Outlier, Change and Burst Simultaneously over Data Stream
ICDMW '07: Proceedings of the Seventh IEEE International Conference on Data Mining WorkshopsEvent detection is one of the most important issues of event processing system, especially Complex Event Processing (CEP). Outlier event, change event and burst event are three typical types of event that need to be identified. Current research works ...
Comments