Research on Operation Characteristics of OD Path Flow Based on Vehicle License Plate Data

Vehicle license plate recognition data contains large-scale vehicle spatiotemporal movement information, which can be used to determine important traffic flow characteristics. Considering the recognition data lack of vehicle travel state information, this paper first proposes a decision boundary division method to identify the travel completion time of each OD pair at directed intersections based on Gaussian mixture model (GMM). Identification of a single trip from original vehicle trajectory data which contains multiple trips is realized. Then, based on the shortest path principle and the structured process of time difference algorithm, the spatiotemporal information imputation of the whole travel chain at road network level is completed. Finally, this method is verified by an example dataset of Kunshan. The path flow information between OD pairs at directed intersections is effectively extracted. And the operation characteristics of the path flow between associated intersections are analyzed.