交通信息与安全2013，Vol.31Issue(2)：94-99,6.DOI:10.3963/j.issn1674-4861.2013.02.021
基于语义层次组成的ST-MRF交通事故检测算法
A Spatio-temporal Markov Random Fields(MRF)Algorithm for Incident Detection
摘要
Abstract
Traditional traffic incident detection algorithms do not consider the overlaps between vehicles and therefore often result in low detection accuracy, especially when traffic volume is high. In order to solve this issue, an incident detection algorithm based on semantic hierarchy coupled with the Spatio-temporal MRF (ST-MRF) model is put forward. The algorithm uses the ST-MRF model to track vehicles, monitors vehicle's location and motion vector, calculates the parameters of traffic flow, and uses a semantic hierarchy algorithm to detect traffic incident by combining such parameters with the data of downstream traffic flow. In order to verify the accuracy of the proposed algorithm, this paper compares its performance with that from the algorithm that detects accidents through tracking vehicles with ST-MRF and corresponding traffic flow data. It is found out that the proposed algorithm has a higher accuracy rate. The study results indicate that the proposed algorithm can accurately detect traffic accidents, even when traffic volume is high, and vehicles appear to overlap each other.关键词
事件检测/ST-MRF/车辆跟踪/语义层次/目标地图/运动矢量Key words
incident detection/ ST-MRF/ vehicle tracking/ semantic hierarchy/ object-map/ motion vector分类
交通工程引用本文复制引用
周君,程琳..基于语义层次组成的ST-MRF交通事故检测算法[J].交通信息与安全,2013,31(2):94-99,6.基金项目
国家自然科学基金项目(批准号:51078085)资助 （批准号:51078085）
