ACCIDENT DETECTION USING VIDEO SURVEILLANCE

Abstract

Computer vision-based accident detection through video surveillance has become a beneficial task. There are many existing technologies to detect accidents in traffic but the speed and accuracy of the system does not meet the requirements. Thus the author formulates the idea of accident detection using video surveillance. The system consists of three phases: Vehicle detection, Vehicle tracking and Accident detection. The proposed framework capitalizes on Mask R-CNN for accurate object detection and uses an efficient centroid-based object tracking algorithm for surveillance footage. The accident detection phase includes Acceleration Anomaly, Trajectory Anomaly, and Change in Angle Anomaly. The probability of an accident is determined by speed and trajectory anomalies in a vehicle after an overlap with other vehicles. If the value is greater than the threshold value, then we can confirm that there is an accident. The proposed framework provides a robust method to achieve a high Detection Rate and a Low False Alarm Rate on general road-traffic CCTV Surveillance footage.