PEDESTRIAN SPEED ESTIMATION METHOD APPLICATION FOR STATISTICAL ANALYSIS OF CROWD DYNAMICS
DOI:
https://doi.org/10.32689/maup.it.2025.1.34Keywords:
сomputer vision, object tracking, gait speed estimation, video surveillance, deep learning, crowd monitoringAbstract
This paper addresses the problem of pedestrian detection, tracking, and gait speed estimation based on video footage from a surveillance camera positioned above the walking area. Such a configuration is typical for surveillance systems in public spaces. The method proposed by the authors does not require any specialized equipment, making it suitable for a wide range of real-world scenarios. The aim of this study is to evaluate the previously developed method for pedestrian speed estimation in a real-life case – video surveillance in a school corridor. The research seeks to assess the effectiveness and reliability of the algorithm under conditions that differ from a controlled laboratory environment. Research methodology. To test the method, a video recording was conducted in a school corridor, where a surveillance camera captured pedestrian flows over a period of 6.5 hours. Using computer vision algorithms, pedestrians were detected, their trajectories tracked, and walking speeds estimated. In total, 1,841 trajectories were extracted, with an average walking speed of 1.15 m/s. The collected data enabled a statistical analysis that revealed patterns of pedestrian traffic under both normal and emergency conditions. Scientific novelty. The study presents the first real-world validation of the previously proposed method, implemented without any prior environmental preparation or the use of additional sensors. A distinguishing feature of the experiment is the consideration of external factors, particularly regular air raid alerts occurring during wartime in Ukraine. Their impact on pedestrian behavior was identified and reflected in the analytical results. This approach contributes to a deeper understanding of human adaptive behavior in stressful situations. Conclusions. The conducted study confirmed the effectiveness of the previously developed method for pedestrian detection, tracking, and speed estimation in real-world conditions. The obtained results can be used to further improve decision support systems in the fields of safety, evacuation planning, and crowd behavior analysis. The use of real data and the inclusion of a crisis context significantly enhance the practical value of the proposed methodology.
References
A. Boroomand, G. Shaker, P. P. Morita, A. Wong and J. Boger, “Autonomous gait speed estimation using 24GHz FMCW radar technology,” 2018 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI), Las Vegas, NV, USA, Apr 2018, pp. 66–69. https://doi.org/10.1109/BHI.2018.8333371.
B. Kannadasan and K. Yogeswari, “ANN-based traffic planning model using adaptive YOLOv5-aided object counting from surveillance videos”. Asian Journal of Civil Engineering, vol. 25, pp. 4771–4788, May 2024. https://doi.org/10.1007/s42107-024-01079-9.
B. Y. Wang, A. Bochkovskiy, H. Y. M. Liao. “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,” In CVPR, New Orleans, LA, USA, 2022, pp. 7464–7475. https://doi.org/10.48550/arXiv.2207.02696.
DBN V.1.1-7:2016, Fire Safety of Building Structures, Ministry of Regional Development of Ukraine, Kyiv, 2016. URL: https://e-construction.gov.ua/files/new_doc/3019922918881626107/2023-01-20/5c73ab5e-fc7d-4433-8bcf-6f78978113b6.pdf.
H. Ghomashchi, J. Paterson, A. C. Novak, and T. Dutta, “Estimating pedestrian walking speed at street crossings using the YOLOv4 and deep SORT algorithms: Proof of principle,” Applied Ergonomics, vol. 119, Article 104292, Apr. 2024. https://doi.org/10.1016/j.apergo.2024.104292.
M. F. M. Ali, M. S. Abustan, et al., “A Case Study of Malaysian Pedestrian Walking Speed at Shopping Malls in Kuala Lumpur, Malaysia using Human Behaviour Simulator (HBS),” International Journal of Integrated Engineering, vol. 11, no. 4, pp. 225–230, Aug. 2019. https://doi.org/10.30880/ijie.2019.11.04.025.
M. Giannoulaki, Z. Christoforou, “Pedestrian Walking Speed Analysis: A Systematic Review,” Sustainability, vol. 16, no. 11, Article 4813, Jun. 2024. URL: https://translate.google.com/website?sl=en&tl=uk&hl=uk&prev=search&u=https://doi.org/10.3390/su16114813.
M. Tishkov. Pedestrian speed estimation. 2024. URL: https://github.com/MaksymTishkov/pedestrian_speed_estimation.
M. W. Park, S. G. Chung, J. Beom, K. S. Kim, J. Kim, and C. H. Park, “Minimum Required Distance for Clinically Significant Measurement of Habitual Gait Speed,” Research Square, 2024. https://doi.org/10.21203/rs.3.rs-5403357/v1.
O. L. Tymoshchuk, M. O. Tishkov, V. G. Bondarenko, “Crowd navigation monitoring during emergencies,” Systemresearch and information technologies, no. 4, pp. 43–54, Dec. 2024. https://doi.org/10.20535/SRIT.2308-8893.2024.4.03.
P. Shrivastav and A. K. T. Andu, “Integrated Approach for Real-time Human Counting, Tracking, and Direction Estimation using Advanced Algorithms,” 2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT), Kamand, India, Nov. 2024, pp. 1–6. https://doi.org/10.1109/ICCCNT61001.2024.10726257.
S. Mirzaei, “Dual-Camera Intersection Monitoring: Detection, Tracking, and Predictive Safety Alerts for Road Users,” M.A.Sc. thesis, Dept. of Mechanical Eng., McMaster Univ., 2024. http://hdl.handle.net/11375/30193.
X. Peng and J. Shan, “Detection and tracking of pedestrians using Doppler LiDAR,” Remote Sensing, vol. 13, no. 15, Article 2952, Jul. 2021. https://doi.org/10.3390/rs13152952.
Y. Liu, W. Zhang, L. Cao and D. Wei, “Ped-Mag-ODO: Indoor Pedestrian Motion Speed Estimation Method Based on Dual Magnetometers,” IEEE Internet of Things Journal, Dec. 2024. https://doi.org/10.1109/JIOT.2024.3512504.