Real-Time Crowd Analysis and Supervised CCTV System Using YOLOv5

The number of people in places like malls, railway stations, airports, stadiums and events is increasing really fast. This makes it very hard to keep everyone safe and watch what is happening all the time. The old CCTV systems need someone to watch them all the time, which's not easy for security people to do. They have to look at videos at the same time and find anything suspicious or if there are too many people in one place. This is why we need a system that can watch the crowd and keep people safe without needing someone to always be watching. This project is about making a system called "Real-Time Crowd Analysis and Supervised CCTV System Using YOLOv5". This system uses computer programs to find people track where they are going see how crowded it is find anything strange watch live and send alerts on Telegram. It looks at the videos from CCTV cameras and live webcams in time to find people see where they are moving check the crowd and find anything suspicious automatically. The YOLOv5 program is very good at finding people in time. The BotSort program helps keep track of people by giving them an ID. The system can also see if there are many people or if they are behaving strangely like running suddenly. If it finds anything suspicious it sends an alert, on Telegram with a picture of what's happening. This helps security people respond quickly.   The system has an interface called Streamlit that shows the live video, crowd information tracking details and alerts in real time. It is easy to use and understand. When we tested the system it worked well with videos and live webcam input. It found people. Tracked them accurately.   Overall this system is an better way to watch crowds and keep people safe. It uses computer programs to find people track them and send alerts, which makes it easier for security people to do their job. This project shows how computer vision and deep learning can be used to keep people safe and make surveillance systems better. The Real-Time Crowd Analysis and Supervised CCTV System Using YOLOv5 is an example of how technology can be used to improve public safety.

[1] J. Redmon, S. Divvala, R. Girshick and A. Farhadi wrote a paper called "You Look Once: Unified Real-Time Object Detection" in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition which was held in 2016. This paper is on pages 779–788.   [2] A. Bochkovskiy, C.-Y. Wang and H.-Y. M. Liao wrote a paper called "YOLOv4: Speed and Accuracy of Object Detection" in 2020. You can find this paper on the arXiv website.   [3] G. Jocher and his team created a GitHub Repository for YOLOv5 by Ultralytics in 2020. You can find the repository at https://github.com/ultralytics/yolov5.   [4] A. B. Chan, Z.-S. J. Liang and N. Vasconcelos wrote a paper called "Privacy Preserving Crowd Monitoring: Counting People Without People Models or Tracking" in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition which was held in 2008. This paper is on pages 1–7.   [5] R. Mehran, A. Oyama and M. Shah wrote a paper called "Abnormal Crowd Behavior Detection Using Social Force Model" in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition which was held in 2009. This paper is on pages 935–942.   [6] N. Wojke, A. Bewley and D. Paulus wrote a paper called "Simple Online and Realtime Tracking with a Deep Association Metric" in the Proceedings of the IEEE International Conference on Image Processing, which was held in 2017. This paper is on pages 3645–3649.   [7] Y. Zhang and his team wrote a paper called "ByteTrack: -Object Tracking by Associating Every Detection Box" in the Proceedings of the European Conference on Computer Vision, which was held in 2022. This paper is on pages 1–21.   [8] R. Girshick wrote a paper called "Fast R-CNN" in the Proceedings of the IEEE International Conference on Computer Vision, which was held in 2015. This paper is on pages 1440–1448.   [9] B. Yang, R. Cao and X. Zhang wrote a paper called "Real-Time Crowd Density Monitoring Using Deep Learning Techniques" in the International Journal of Computer Vision and Image Processing in 2021. This paper is in volume 12 number 3. Is on pages 45–56.   [10] S. Khan and M. Hussain wrote a paper called "AI-Based Intelligent Surveillance System for Crowd Monitoring and Anomaly Detection" in the International Journal of Advanced Computer Science and Applications in 2022. This paper is in volume 13 number 4. Is on pages 210–218.