Research Of Dynamic Origin-Destination Matrix Estimation Based On Video Vehicle License Plate Recognition

Dynamic OD matrix is the basic data of transportation network planning and traffic operation management. And it is also the basis for the decision making of ITS in traffic control and route guidance programming. Acquirement of dynamic OD matrix is one of the cores of the ITS study. In recent years, we can acquire vehicles license plate through the video image information, on the basis of which, we can calculate the transportation information such as travel time, OD, and so on. In China, this technology is being used in Beijing, Shanghai and Hangzhou. And it is large-scale promoted implementation.However, the present researches pay more attention to how enhance the video vehicle license plate recognition precision. Video vehicle license plate recognition's layout planning, OD and path correction method haven't been studied deeply. We are lack of methods and tools to do such researches.This paper has initially discussed the framework of 'Research of Dynamic OD Matrix Estimation Based on Video Vehicle License Plate Recognition'. In this paper, I proposed the research technique which is based on the simulation test platform. I used VISSIM, Microsoft Access and Visual the Basic as tools to study. Simultaneously, the paper has deeply studied in 'Under the condition which is building video vehicle license plate recognition on every lane and entrance of the study area, how examination precision will influence the dynamic OD matrix acquirement'.The main research of this paper is as follows. (1) I built the simulation test platform, which has widespread application in 'Research of Dynamic OD Matrix Estimation Based on Video Vehicle License Plate Recognition'. (2) In order to enhance dynamic OD matrix and path recognition precision, I researched three different kinds of calculation method. (3) I proposed three OD precision evaluation indexes. (4) I used optimized algorithm to do the sensitive analysis. Through the analysis, I established a limit accuracy of the video vehicle license plate recognition system. The limit accuracy of the video vehicle license plate recognition system is 80% precision.