基于出租车轨迹数据的交通异常识别算法研究

为了实现基于出租车轨迹数据的交通异常识别,本文首先以城市栅格地图模型为框架,提出了一种针对城市路网的多光谱分隔算法,并根据城市路网分别从区域增长与区域融合两种角度实现了多光谱地图的分割。其次在分割的城市路网基础上,设计了交通异常的识别算法。算法依据单元区域内道路网络拓扑结构构建交通异常图,然后根据出租车路径选择模式的历史规律计算每个单元区域内不同路径上的出租车轨迹流量的变化,最后根据三倍均方差指标识别单元区域内的交通异常。文章最后以哈尔滨为例进行了算例分析,算例结果表明,本文提出的异常识别算法取得了良好的效果,验证了算法的有效性及准确性。

The Study of Traffic Anomaly Recognition Based on Taxi Track Date

In order to realize the traffic anomaly recognition based on the taxi track data, this paper takes the urban grid map model as the framework firstly, proposes a multi spectral separation algorithm for urban road network, and realizes the multi spectral map segmentation according to the urban road network from two angles of regional growth and regional integration. Secondly, based on the segmentation of urban road network, a traffic anomaly recognition algorithm is designed. According to the road network topology in the unit area, the algorithm constructs the traffic anomaly map, and then calculates the change of the taxi path flow on the different paths in each unit area according to the history law of the taxi path selection model. The traffic anomalies in the unit area are identified based on the three times mean square error index. Finally, a complete urban road network traffic anomaly map is constructed. At the end of this paper, the example of Harbin is taken as an example. The results show that the proposed anomaly recognition algorithm has achieved good results and verifies the effectiveness and accuracy of the algorithm.