Accident and Crime Detection Using YOLO from Surveillance Video Simulation

The project “Accident and Crime Detection from Surveillance Video Simulation Using YOLO” was developed with the purpose of improving traditional surveillance systems by introducing Artificial Intelligence and Computer Vision technologies. Nowadays, the increasing number of vehicles, population growth, and rise in criminal activities have made public safety an important concern. In many places such as roads, shopping malls, banks, railway stations, and public areas, CCTV cameras are installed for security purposes. However, these cameras continuously generate a large amount of video data that must be monitored by humans. Continuous observation of multiple camera feeds is difficult because a person can become tired, lose concentration, or miss important events. As a result, accidents and criminal activities may not be identified immediately, causing delays in taking action. The proposed system attempts to solve this problem by creating an intelligent surveillance system that can automatically monitor video footage and detect incidents in real time. In this project, the YOLO (You Only Look Once) deep learning model was used because of its fast processing speed and efficient object detection capability. The system continuously analyzes surveillance videos, identifies objects such as vehicles and people, and determines whether any abnormal activity is taking place. During the implementation and testing of the project, the system successfully detected different events such as vehicle accidents, suspicious human behavior, crowd-related situations, and crime-related activities. Whenever an unusual event was detected, the system generated alerts and stored information in the database. The dashboard also displayed live monitoring information, incident history, and related details, making it easier for users to understand and analyze events. The project showed that Artificial Intelligence can significantly reduce human effort in surveillance monitoring. Instead of depending completely on manual observation, intelligent systems can automatically identify important incidents and assist authorities in making quicker decisions. This can help reduce response time during emergencies and improve public safety. Although the project achieved its intended objectives, some limitations were observed during testing. Factors such as poor video quality, low lighting conditions, crowded environments, and object overlap sometimes affected the detection accuracy. Despite these limitations, the overall results were satisfactory and proved that intelligent video analysis systems can be effectively used in surveillance applications.

[1]. Sanjeev Kukreti; RVS Praveen; Saloni Bansal; Hemanth Raju; Navdeep Singh; Rowsonara Begum “Object Detection in Real-Time Surveillance Using Deep Learning-Based YOLO Framework” in 2025 International Conference on Computational, Communication and Information Technology (ICCCIT), February 2025 [2]. Varad Agarwal, Yuvraj Singh Rathore, Shyam Singh, Indra Kishor “Smart Surveillance System Using AI for Anomaly Detection” in 2025 International Journal of Innovative Research in Technology (IJIRT), Volume 11 Issue 11 Apr 2025. [3]. Manu, Y. M., S. V. Shashikala, N. Darshan, A. P. Dheeraj, N. J. Hemanth, and B. Nishanth Gowda. "Smart Surveillance Camera Using AI." In 2024 Second International Conference on      Advances      in       Information Technology (ICAIT), vol. 1, pp. 1-6. IEEE, 2024.  [4]. Maheshwari, K.K., Hajano, S., Baloch, S., Zafar, Z. and Fatima, T., 2025, April. AI Based Surveillance System. In 2025 International Conference on Emerging Technologies in Electronics, Computing, and Communication (ICETECC) (pp. 1-6). IEEE. [5]. Villegas-Ch, W., García-Ortiz, J. and Sánchez-Viteri, S., 2024. Toward intelligent monitoring in IoT: AI applications for real-time analysis and prediction. IEEE Access, 12, pp.40368-40386. [6]. Siddiqui, Muhammad Zulqarnain, Kamran Khan, and Abdul Jameel Khan. "Smart Surveillance and Crime Control Using an AI-Based Model." Journal of Computing & Biomedical Informatics 6, no. 01 (2023): 216-221. [7]. Kakadiya, R., Lemos, R., Mangalan, S., Pillai, M. and Nikam, S., 2019, June. Ai based automatic robbery/theft detection using smart surveillance in banks. In 2019 3rd International conference    on  Electronics, Communication and Aerospace Technology (ICECA) (pp. 201-204). IEEE. [8]. Saranya E, Vanitha A, Selvanandhini A, Soundarya VN, Vinisha M. A Survey on AI Based Smart CCTV Surveillance  System.  In2024  10th International Conference on Advanced Computing and Communication Systems (ICACCS) 2024 Mar 14 (Vol. 1, pp. 1-5). IEEE.  [9]. Rahman, Md Abdur, M. Shamim Hossain, Ahmad J. Showail, Nabil A. Alrajeh, and Ahmed Ghoneim. "AI-enabled IIoT for live smart city event monitoring." IEEE Internet of Things Journal 10, no. 4 (2021): 2872-2880. [10].           Qaraqe M, Elzein A, Basaran E, Yang Y, Varghese EB, Costandi W, Rizk J, Alam N. PublicVision: A secure smart surveillance system for crowd behavior recognition. IEEE Access. 2024 Feb 16;12:26474-91.