Crowd Detection and Overcrowding Alert System

Abstract

A crowd detection and overcrowding alert system is proposed to monitor public spaces and prevent unsafe crowd conditions. The system applies computer vision and deep learning techniques to analyze live video feeds from surveillance cameras, detecting the number of people present, tracking movement patterns, and estimating crowd density in specific zones. When density exceeds a predefined safety threshold, the system automatically generates alerts for authorities, enabling timely intervention and helping to prevent accidents, stampedes, and congestion in locations such as malls, transportation hubs, campuses, and public events. The proposed solution is designed to integrate with existing CCTV infrastructure as well as IoT devices and mobile notification services for faster response, and to provide continuous monitoring and visual insight through a live dashboard so that authorities can easily identify high-risk zones. By reducing dependence on manual supervision, the system is intended to increase surveillance efficiency, support scalable deployment in both small and large environments, and lay the groundwork for future extensions such as behavior analysis and anomaly detection, contributing to public safety and smart-city crowd management.

Introduction

The proposed Crowd Detection and Overcrowding Alert System aims to improve public safety by automatically detecting people, estimating crowd density, and generating real-time alerts when overcrowding exceeds predefined safety limits. Traditional crowd monitoring relies on manual CCTV surveillance, which is time-consuming, prone to human error, and inefficient for monitoring large public spaces such as railway stations, airports, shopping malls, stadiums, and public gatherings.

The system uses computer vision and deep learning techniques, particularly the YOLOv8 object detection model, to identify and count people in live or recorded video streams. It processes video frames, detects individuals, estimates crowd density within specific regions, and compares the people count against predefined safety thresholds. When the threshold is exceeded, an automatic alert is generated to notify authorities and support timely intervention.

The literature survey examines existing approaches based on YOLO models, convolutional neural networks (CNNs), optical flow, LSTM networks, object tracking, and anomaly detection. Although these techniques have demonstrated promising results in crowd counting and abnormal behavior recognition, many existing systems face limitations such as reduced accuracy in extremely dense crowds, high computational requirements, camera-specific thresholds, and insufficient real-time alerting capabilities.

The proposed methodology consists of five main stages: video acquisition and preprocessing, person detection using YOLOv8, crowd density estimation, threshold comparison and alert generation, and dashboard-based monitoring. The dashboard displays the live people count, crowd status, and alerts for monitored areas. Object tracking and motion analysis can also help identify crowd movement patterns and may be extended to recognize abnormal activities such as sudden rushing or pushing.

Conclusion

This paper has presented the design of a Crowd Detection and Overcrowding Alert System that combines computer vision and deep learning to detect people, estimate crowd density, and issue automatic alerts when a predefined safety threshold is exceeded. By building on existing CCTV infrastructure and a YOLOv8-based detection pipeline, the proposed approach aims to reduce reliance on manual surveillance, support faster response through dashboard and notification integration, and provide a scalable foundation for future enhancements such as behavior analysis, anomaly detection, and predictive crowdflow modeling in smart-city environments.

References

[1] D. Mabrouk, M. A. Abdel-Fattah, and A. Taha, “Robust crowd anomaly detection via hybrid ensemble learning for real-world surveillance,” Scientific Reports, vol. 15, Nov. 2023. DOI: 10.1038/s41598-025-27408-9 [2] M. S. Gündüz and G. I?ik, “A new YOLO-based method for real-time crowd detection from video and performance analysis of YOLO models,” Journal of Real-Time Image Processing, vol. 20, no. 1, p. 5, Jan. 2023. DOI: 10.1007/s11554-023-01276-w [3] K. Rezaee, S. M. Rezakhani, M. R. Khosravi, and V. G. Sangeetha, “A survey on deep learning-based real-time crowd anomaly detection for secure distributed video surveillance,” Personal and Ubiquitous Computing, vol. 28, pp. 135–151, 2024. DOI: 10.1007/s00779-021-01586 [4] M. Mehrabi et al., “A cloud-based deep learning framework for early detection of pushing at crowded event entrances,” arXiv preprint, arXiv:2302.08237, Feb. 2025. [5] J. Kim et al., “A real-time crowd density level detection system,” in Communications in Computer and Information Science, vol. 1978, Springer, Cham, Dec. 2025, pp. 145–157. DOI: 10.1007/978-3-031-49529-8_12 [6] M. H. Sharif, L. Jiao, and C. W. Omlin, “Deep crowd anomaly detection: state-of-the-art, challenges, and future research directions,” Artificial Intelligence Review, vol. 58, Feb. 2025. DOI: 10.1007/s10462-024-11092-8 [7] K. Yasen, J. Zhou, N. Zhou, K. Qin, Z. Wang, and Y. Li, “Crowd gathering detection method based on multi-scale feature fusion and convolutional attention,” Sensors, vol. 25, no. 21, p. 6550, Oct. 2025. DOI: 10.3390/s25216550 [8] S. Vasavi, C. Neeraja, and A. Kruthika, “Crowd Density Analysis for Public Safety Using YOLOv8,” in 2025 12th International Conference on Reliability, Infocom Technologies and Optimization (ICRITO), Noida NCR, India, 2025, pp. 1–5. DOI: 10.1109/ICRITO66076.2025.11241798. [9] R. Jayachitra, R. Ragul, and C. R. Krishna, “Next-Generation Public Safety System for Real-Time Crowd and Weapon Detection using YOLOv11 with Automated Alerts,” in 2025 3rd International Conference on Advancement in Computation & Computer Technologies (InCACCT), Gharuan, India, 2025, pp. 767–772. DOI: 10.1109/InCACCT65424.2025.1101140. [10] A. Kumar, A. Shukla, P. Garg, S. Sahu, and P. Shukla, “Leveraging Machine Learning for Real-Time Crowd Control and Safety at Kumbh Mela,” in 2025 3rd International Conference on Communication, Security, and Artificial Intelligence (ICCSAI), Greater Noida, India, 2025, pp. 1416–1422. DOI: 10.1109/ICCSAI64074.2025.11064144.

Copyright

Copyright © 2026 Mrs. Ayesha Siddiqa , Chaitra P, Varsha H R, Sahana H V, Yashaswini S N. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.