Smart CCTV Monitoring and Automated Emergency Alert Systems

Author

I. A. Jannathul Firthous , Kannika Devi P

Abstract

 Traditional closed-circuit television (CCTV) systems depend on human operators to watch live feeds or review footage after an event, which is difficult when many cameras operate at once. This paper presents an AI-based smart CCTV surveillance prototype that analyzes video to help identify potential violence, dangerous objects, and suspicious activity. OpenCV reads video streams and prepares frames, a YOLO detector identifies people and selected weapon classes, a tracking algorithm associates detections across frames, and a CNN-LSTM action-recognition model is proposed for violence/non-violence classification. A configurable decision layer flags potential incidents, after which the application saves evidence frames or clips and stores incident metadata in a MySQL database. A Flask backend and web dashboard let authorized personnel review incidents, evidence, timestamps and alert status, while e-mail notifications draw attention to important events sooner. The system is designed as a decision-support tool, not an autonomous authority: AI predictions may be wrong, and human review is required before a flagged event is treated as confirmed misconduct. Performance is to be assessed on a held-out dataset using precision, recall, mAP, F1-score and inference speed.

Keywords

Smart CCTV; Computer Vision; YOLO; Object Detection; Action Recognition; CNN-LSTM; Person Tracking; Incident Reporting; MySQL; Flask; Automated Alerts

Full Text:

Download Paper PDF

References

Carreira, J., & Zisserman, A. (2017). Quo vadis, action recognition? A new model and the Kinetics dataset. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 6299–6308. https://doi.org/10.1109/CVPR.2017.502

Cheng, M., Cai, K., & Li, M. (2021). RWF-2000: An open large scale video database for violence detection. Proceedings of the 25th International Conference on Pattern Recognition (ICPR), 4183–4190. https://doi.org/10.1109/ICPR48806.2021.9412502

Donahue, J., Hendricks, L. A., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., & Darrell, T. (2015). Long-term recurrent convolutional networks for visual recognition and description. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2625–2634. https://doi.org/10.1109/CVPR.2015.7298878

Donahue, J., Hendricks, L. A., Guadarrama, S., Rohrbach, M., Venugopalan, S., Saenko, K., & Darrell, T. (2017). Long-term recurrent convolutional networks for visual recognition and description. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(4), 677–691. https://doi.org/10.1109/TPAMI.2016.2599174

Kay, W., et al. (2017). The Kinetics human action video dataset. arXiv preprint arXiv:1705.06950. https://arxiv.org/abs/1705.06950

OpenCV. (n.d.). OpenCV documentation. https://docs.opencv.org/

Oracle. (n.d.). MySQL reference manual. https://dev.mysql.com/doc/

OWASP Foundation. (n.d.). Password storage cheat sheet. https://cheatsheetseries.owasp.org/cheatsheets/Password_Storage_Cheat_Sheet.html

Pallets Projects. (n.d.). Flask documentation. https://flask.palletsprojects.com/

PyTorch. (n.d.). PyTorch documentation. https://pytorch.org/docs/

Sultani, W., Chen, C., & Shah, M. (2018). Real-world anomaly detection in surveillance videos. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 6479–6488. https://doi.org/10.1109/CVPR.2018.00678

Ultralytics. (n.d.). YOLO11. Ultralytics documentation. https://docs.ultralytics.com/models/yolo11/

Share your valuable work from Social Media Buttons