An Improved Object Detection and Trajectory Prediction Method for Traffic Conflicts Analysis

Lu Yang; Ahmad Sufril Azlan Bin Mohamed; Majid Khan Bin Majahar Ali

doi:10.7307/ptt.v35i4.173

Authors

Lu Yang School of Computer Sciences, Universiti Sains Malaysia
Ahmad Sufril Azlan Bin Mohamed School of Computer Sciences, Universiti Sains Malaysia
Majid Khan Bin Majahar Ali School of Mathematical Sciences, Universiti Sains Malaysia

DOI:

https://doi.org/10.7307/ptt.v35i4.173

Keywords:

near-miss, object detection, object tracking, trajectory prediction

Abstract

Although computer vision-based methods have seen broad utilisation in evaluating traffic situations, there is a lack of research on the assessment and prediction of near misses in traffic. In addition, most object detection algorithms are not very good at detecting small targets. This study proposes a combination of object detection and tracking algorithms, Inverse Perspective Mapping (IPM), and trajectory prediction mechanisms to assess near-miss events. First, an instance segmentation head was proposed to improve the accuracy of the object frame box detection phase. Secondly, IPM was applied to all detection results. The relationship between them is then explored based on their distance to determine whether there is a near-miss event. In this process, the moving speed of the target was considered as a parameter. Finally, the Kalman filter is used to predict the object's trajectory to determine whether there will be a near-miss in the next few seconds. Experiments on Closed-Circuit Television (CCTV) datasets showed results of 0.94 mAP compared to other state-of-the-art methods. In addition to improved detection accuracy, the advantages of instance segmentation fused object detection for small target detection are validated. Therefore, the results will be used to analyse near misses more accurately.

References

Loo BPY, Huang Z. Delineating traffic congestion zones in cities: An effective approach based on GIS. J Transp Geogr. 2021;94(6):103108. DOI:10.1016/j.jtrangeo.2021.103-108.

Kitamura Y, Hayashi M, Yagi E. Traffic problems in Southeast Asia featuring the case of Cambodia’s traffic accidents involving motorcycles. IATSS Res. 2018;42(4):163-170. DOI: 10.1016/j.iatssr.2018.11.001.

Zwetsloot G, Leka S, Kines P. Vision zero: From accident prevention to the promotion of health, safety and well-being at work. Policy and Practice in Health and Safety. 2017;15:88-110. DOI: 10.1080/14773996.2017.1308701.

Zheng L, Sayed T, Mannering F. Modeling traffic conflicts for use in road safety analysis: A review of analytic methods and future directions. Anal Methods Accid Res. 2021;29:100-142. DOI: 10.1016/j.amar.2020.100142.

Ge J, et al. Accident causation models developed in China between 1978 and 2018: Review and comparison. Saf Sci. 2022;148(12). DOI:10.1016/j.ssci.2021.105653.

Heinrich HW, Stone RW. Industrial accident prevention. Soc Serv Rev. 1931;5(2):323-324. DOI: 10.1086/630904.

Terum JA, Svartdal F. Lessons learned from accident and near-accident experiences in traffic. Saf Sci. 2019;120(6):672-678. DOI: 10.1016/j.ssci.2019.07.040.

Wu LS, Dong Y, Wu. A case study of road traffic accidents based on Heinrich's accident causation theory. Chinese Journal of Ergonomics. 2018;24(2):60-64. DOI: 10.13837/j.issn.1006-8309.2018.02.0012.

Hajjami LE, Mellouli EM, Berrada M. Neural network based sliding mode lateral control for autonomous vehicle. 2020 1st International Conference on Innovative Research in Applied Science, Engineering and Technology (IRASET). 2020. p. 1-6. DOI: 10.1109/IRASET48871.2020.9092055.

Lhoussain EH, Mellouli EM, Berrada M. Robust adaptive non-singular fast terminal sliding-mode lateral control for an uncertain ego vehicle at the lane-change maneuver subjected to abrupt change. Int. J. Dynam. Control. 2021;9:1765-1782. DOI: 10.1007/s40435-021-00771-x.

Lhoussain EH, et al. A robust intelligent controller for autonomous ground vehicle longitudinal dynamics. Applied Sciences. 2023;13(1):501. DOI: 10.3390/app13010501.

Kataoka H, et al. Drive video analysis for the detection of traffic near-miss incidents. Proc. - IEEE Int. Conf. Robot. Autom. 2018. p. 3421-3428. DOI: 10.1109/ICRA.2018.8460812.

Suzuki T, Aoki Y, Kataoka H. Pedestrian near-miss analysis on vehicle-mounted driving recorders. Proc. 15th IAPR Int. Conf. 2017. p. 416-419. DOI: 10.23919/MVA.2017.7986889.

Ospina MH, et al. Extraction of decision rules using genetic algorithms and simulated annealing for prediction of severity of traffic accidents by motorcyclists. J. Ambient Intell. Humaniz. Comput. 2021(12):10051-10072. DOI: 10.1007/s12652-020-02759-5.

Uchida N, Kawakoshi M, Tagawa T, Mochida T. An investigation of factors contributing to major crash types in Japan based on naturalistic driving dana. IATSS Res. 2010;(34):22-30. DOI: 10.1016/j.iatssr.2010.07.002.

Chen W, Qiao Y, Li Y. Inception-SSD: An improved single shot detector for vehicle detection. J. Ambient Intell. Humaniz. Comput. 2020. DOI: 10.1007/s12652-020-02085-w.

Yang B, et al. A vehicle tracking algorithm combining detector and tracker. Eurasip J. Image Video Process. 2020;(1). DOI: 10.1186/s13640-020-00505-7.

Wan J. An efficient small traffic sign detection method based on YOLOv3. J. Signal Process. Syst. 2020. DOI: 10.1007/s11265-020-01614-2.

Tran AC, et al. A model for real-time traffic signs recognition based on the YOLO algorithm – A case study using vietnamese traffic signs. Futur. Data Secur. Eng. 2019;11814:104-116. DOI: 10.1007/978-3-030-35653-8_8.

Kurniawan J, Dewa CK, Afiahayati. Traffic congestion detection: Learning from CCTV monitoring images using convolutional neural network. Procedia Comput Sci. 2018;144:291-297. DOI: 10.1016/j.procs.2018.10.530.

Abdel-Aty M, Wu Y, Zheng O, Yuan J. Using closed-circuit television cameras to analyze traffic safety at intersections based on vehicle key points detection. Accid Anal Prev. 2022;176:106794. DOI: 10.1016/j.aap.2022.106794.

Bertozzi M, Broggi A, Fascioli A. Stereo inverse perspective mapping: Theory and applications. Image Vis Comput. 1998;16(8):585-590. DOI: 10.1016/s0262-8856(97)00093-0.

Zhao ZQ, Zheng P, Xu ST, Wu X. Object detection with deep learning: A Review. IEEE Trans Neural Networks Learn Syst. 2019;30(11):3212-3232. DOI: 10.1109/TNNLS.2018.2876865.

Galoogahi HK, Sim T, Lucey S. Correlation filters with limited boundaries. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit. 2015;7(6):4630-4638. DOI: 10.1109/CVPR.2015.7299094.

Hurtik P, et al. Poly-YOLO: Higher speed, more precise detection and instance segmentation for YOLOv3. Neural Comput Appl. 2022;34(10):8275-8290. DOI: 10.1007/s00521-021-05978-9.

Zhang Q, Chang X, Bian SB. Vehicle-damage-detection segmentation algorithm based on improved Mask RCNN. IEEE Access. 2020;8:6997-7004. DOI: 10.1109/ACCESS.2020.2964055.

Khan MA, Akram T, Zhang YD, Sharif M. Attributes based skin lesion detection and recognition: A mask RCNN and transfer learning-based deep learning framework. Pattern Recognit Lett. 2021;143:58-66. DOI: 10.1016/j.patrec.2020.12.015.

Wang CY, Bochkovskiy A, Liao H-YM. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. 2022. http://arxiv.org/abs/2207.02696.

Pham V, Pham C, Dang T. Road damage detection and classification with Detectron2 and Faster R-CNN. Proc - 2020 IEEE Int Conf Big Data, Big Data 2020. 2020. p. 5592-5601. DOI: 10.1109/BigData50022.2020.9378027.

Luo W, et al. Multiple object tracking: A literature review. Artif Intell. 2021;293:1-49. DOI: 10.1016/j.artint.2020.103448.

Mangalam K, et al. It is not the journey but the destination: Endpoint conditioned trajectory prediction. Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics). 2020;12347:759-776. DOI: 10.1007/978-3-030-58536-5_45.

Galoogahi HK, Fagg A, Lucey S. Learning background-aware correlation filters for visual tracking. Proc IEEE Int Conf Comput Vis. 2017;2017(10):1144-1152. DOI: 10.1109/ICCV.2017.129.

Wojke N, Bewley A, Paulus D. Simple online and realtime tracking with a deep association metric. 2017 IEEE International Conference on Image Processing (ICIP). 2017. p. 3645-3649. DOI: 10.1109/ICIP.2017.8296962.

Du Y, Song Y, Yang B, Zhao Y. StrongSORT: Make DeepSORT great again. 2022. http://arxiv.org/abs/2202.13514.

Luo H, et al. A strong baseline and batch normalization neck for deep person re-identification. IEEE Trans Multimed. 2020;22(10):2597-2609. DOI: 10.1109/TMM.2019.2958756.

Bruls T, Porav H, Kunze L, Newman P. The right (angled) perspective: Improving the understanding of road scenes using boosted inverse perspective mapping. IEEE Intell Veh Symp Proc. 2019;2019(6):302-309. DOI: 10.1109/IVS.2019.8814056.

Tanveer MH, Sgorbissa A. An inverse perspective mapping approach using monocular camera of pepper humanoid robot to determine the position of other moving robot in plane. ICINCO 2018 - Proc 15th Int Conf Informatics Control Autom Robot. 2018;2(Icinco):219-225. DOI: 10.5220/0006930002190225.

Ivanovic B, Pavone M. Rethinking trajectory forecasting evaluation. ArXiv. abs/2107.10297. DOI: 10.48550/arXiv.2107.10297.

Mahmud SMS, Luis FL, Hoque MS, Tavassoli A. Application of proximal surrogate indicators for safety evaluation: A review of recent developments and research needs. IATSS Research. 2017;41(4):153-163. DOI: 10.1016/j.iatssr.2017.02.001.

Uno N, Iida Y, Itsubo S, Yasuhara S. A microscopic analysis of traffic conflict caused by lane-changing vehicle at weaving section. Proceedings of the 13th Mini-EURO Conference-Handling Uncertainty in the Analysis of Traffic and Transportation Systems. 2002. p. 10-13.

Allen BL, Shin TB, Cooper PJ. Analysis of traffic conflicts and collisions. No. HS-025 846, 1978.

Chen QH, et al. Modeling accident risks in different lane-changing behavioral patterns. Analytic Methods in Accident Research. 2021;(30). DOI: 10.1016/j.amar.2021.100159.