A Traffic Accident Recording and Reporting Model at Intersections.pdf

188 IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 8, NO. 2, JUNE 2007 A Traffic Accident Recording and Reporting Model at Intersections Yong-Kul Ki, Member, IEEE, and Dong-Young Lee Abstract—In this paper, we suggested a vision-based traffic accident detection algorithm and developed a system for auto- matically detecting, recording, and reporting traffic accidents at intersections. A system with these properties would be beneficial in determining the cause of accidents and the features of an intersec- tion that impact safety. This model first extracts the vehicles from the video image of the charge-couple-device camera, tracks the moving vehicles (MVs), and extracts features such as the variation rate of the velocity, position, area, and direction of MVs. The model then makes decisions on the traffic accident based on the extracted features. In a field test, the suggested model achieved a correct detection rate (CDR) of 50% and a detection rate of 60%. Considering that a sound-based accident detection system showed a CDR of 1% and a DR of 66.1%, our result is a remarkable achievement. Index Terms—Accident detection at an intersection, accident recording and reporting system (ARRS), moving picture, traffic accident at an intersection, vehicle tracking. I. INTRODUCTION EVERY year, vehicular accidents cause tragic loss of lives,cost many countries tremendous amount of money, and produce substantial congestion to a nation’s transportation sys- tem. Approximately 50%–60% of the delays on urban freeways are associated with incidents, and on urban surface streets, a large percentage of traffic accidents and most delays occur at or near intersections [1]. Intersections are a common place for crashes, which may be due to the fact that there are several conflicting movements, as well as a myriad of different intersec- tion design characteristics. Intersections also tend to experience severe crashes due to the fact that several types of injurious crashes, such