Prasad Toshniwal, Pankaj Jadhav, Namrata Rajbhar, Seema Rode

DOI: https://doi.org/10.5281/zenodo.24197101

Road traffic accidents require rapid detection and reliable communication of incident information to emergency responders. Conventional reporting often depends on witnesses or manual observation, while automated solutions based only on video or vehicle sensors can fail under difficult conditions or generate false alerts. This literature review examines research on computer-vision-based accident detection, deep-learning object detection, vehicle tracking, Internet of Things (IoT) sensing, automatic license plate recognition (ALPR), and lightweight communication. Studies on traffic-video anomaly detection show that spatiotemporal models can identify unusual events, while trajectory-based methods can detect conflicts among road users. ALPR surveys identify illumination, motion blur, occlusion, camera position, and regional plate variation as persistent challenges. These findings support a multimodal approach in which video analytics are combined with impact and motion sensors, GPS coordinates, vehicle identification, and a communication layer such as MQTT. Edge processing may reduce dependence on continuous cloud round trips, while cloud services can support centralized records and monitoring. However, the reviewed literature does not establish that one integrated pipeline will automatically improve every metric. Further evaluation is required for false-positive rate, detection latency, severity classification, plate-recognition performance, communication reliability, privacy, and emergency-service integration.

Road accident detection; computer vision; Internet of Things; YOLO; sensor fusion; emergency response; vehicle tracking; MQTT.

This literature review examined technical foundations for AI-based road accident detection and emergency response using computer vision and IoT. Video anomaly detection and trajectory-conflict analysis show how temporal behavior can support accident-candidate detection [1], [2]. Object detection and tracking provide useful components, but they do not independently prove that a collision has occurred [3], [4], [7], [8]. ALPR may contribute vehicle identification, although image conditions and plate variation can reduce reliability [9], [10]. MQTT provides a lightweight communication mechanism, but dependable emergency notification also requires application-level security, acknowledgement, retries, duplicate handling, and operational authorization [5]. The review therefore motivates a modular pipeline combining video analytics, optional sensor evidence, confidence-aware vehicle identification, GPS localization, secure messaging, and human oversight. Its practical value remains to be established through reproducible tests of detection quality, false alarms, latency, communication reliability, privacy, and usability. The proposed architecture should be treated as a design direction for future implementation, not as an experimentally validated system.

[1] K. Pawar and V. Attar, “Deep learning based detection and localization of road accidents from traffic surveillance videos,” ICT Express, vol. 8, no. 3, pp. 379–387, 2022. doi: 10.1016/j.icte.2021.11.004.
[2] H. Ghahremannezhad, H. Shi, and C. Liu, “Real-Time Accident Detection in Traffic Surveillance Using Deep Learning,” in Proc. IEEE International Conference on Imaging Systems and Techniques (IST), 2022. doi: 10.1109/IST55454.2022.9827736.
[3] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788. doi: 10.1109/CVPR.2016.91.
[4] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” arXiv:2004.10934, 2020.
[5] OASIS, MQTT Version 5.0, OASIS Standard, Mar. 2019. Available: https://docs.oasis-open.org/mqtt/mqtt/v5.0/mqtt-v5.0.html
[6] G. Jocher et al., “Ultralytics YOLO,” software documentation and repository. Available: https://github.com/ultralytics/ultralytics. Accessed Oct. 12, 2026.
[7] A. Bewley, Z. Ge, L. Ott, F. Ramos, and B. Upcroft, “Simple Online and Realtime Tracking,” in Proc. IEEE International Conference on Image Processing (ICIP), 2016, pp. 3464–3468. doi: 10.1109/ICIP.2016.7533003.
[8] N. Wojke, A. Bewley, and D. Paulus, “Simple Online and Realtime Tracking with a Deep Association Metric,” in Proc. RESEARCH PROJECT GROUP 36 • LITERATURE REVIEW IEEE International Conference on Image Processing (ICIP), 2017, pp. 3645–3649. doi: 10.1109/ICIP.2017.8296962.
[9] J. Shashirangana, H. Padmasiri, D. Meedeniya, and C. Perera, “Automated License Plate Recognition: A Survey on Methods and Techniques,” IEEE Access, vol. 9, pp. 11203–11225, 2021. doi: 10.1109/ACCESS.2020.3047929.
[10]M. A. Khan et al., “Automatic Number Plate Recognition: A Detailed Survey of Relevant Algorithms,” Sensors, vol. 21, no. 9, article 3028, 2021. doi: 10.3390/s21093028.
[11] J. Fang, Y. Qiao, J. Xue, and Z. Li, “Vision-Based Traffic Accident Detection and Anticipation: A Survey,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 4, pp. 1983–1999, Apr. 2024, doi: 10.1109/TCSVT.2023.3307655.
[12] M. A. Sahraei and S. R. M. Al Mamari, “A Review of Internet of Things Approaches for Vehicle Accident Detection and Emergency Notification,” Sustainability, vol. 17, no. 14, Art. no. 6510, 2025, doi: 10.3390/su17146510.
[13] A. Razi and X. Chen, “Deep learning serves traffic safety analysis: A forward-looking review,” IET Intelligent Transport Systems, vol. 17, no. 1, pp. 22–71, 2023, doi: 10.1049/itr2.12257.
[14] A. B. Z. Chai, B. T. Lau, M. K. T. Tee, and C. McCarthy, “Enhancing road safety with machine learning: Current advances and future directions in accident prediction using non-visual data,” Engineering Applications of Artificial Intelligence, vol. 137, Art. no. 109086, 2024, doi: 10.1016/j.engappai.2024.109086.
[15] C. R. Prabu and K. Namitha, “IoT-based automatic road accident detection systems: A review,” in Proc. 4th International Conference on Soft Computing for Security Applications (ICSCSA), 2024, doi: 10.1109/ICSCSA64454.2024.00122

Prasad Toshniwal, Pankaj Jadhav, Namrata Rajbhar, Seema Rode (2026). AI-BASED REAL-TIME ROAD ACCIDENT DETECTION AND EMERGENCY RESPONSE SYSTEM USING COMPUTER VISION AND IoT: A LITERATURE REVIEW. International Journal of Computer Techniques, 13(6), 742–749. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.24197101