Real- Time Accident Detection Using CCTV Footage Analysis

The rapid growth of vehicular traffic has led to an increase in road accidents, emphasizing the need for efficient, real-time detection systems. This paper introduces a novel approach using Convolutional Neural Networks (CNNs) and advanced computer vision techniques to detect accidents through CCTV footage. The system monitors traffic in real time, achieving a high accuracy of 94.3 %, precision of 91.8 %, and recall of 90.5%, with a robust ROC AUC score of 0.97. By analyzing the patterns in the video stream, the system can accurately detect accidents, triggering immediate alerts to respective authorities. The implementation of this system demonstrates its potential as a valuable tool in modern smart city infrastructure, contributing to safer and more efficient urban environments.