Big traffic data processing framework for intelligent monitoring and recording systems

In recent years, intelligent monitoring and recording system (IMRS) has been widely used in many cities. It is a networked system which includes the front-end image acquisition system and the back-end data processing platform. In the front-end acquisition system, after recording the image data the front-end processing system analyzes the image data and automatically extracts some information of the passing vehicles, such as time, location, direction, the number and color of license plate, etc. After that, the information is sent to the back-end data processing platform for deeper analysis, such as vehicle trajectory tracking and traffic state estimation. However, as the data scales to a great amount, it becomes difficult for traditional analysis framework and tools to meet the requirement of big data analysis. HBase is the column-oriented database based on Hadoop, along with Map-Reduce programming paradigm. It can deal with the problems in processing big data that intelligent monitoring and recording system faces nowadays. In this paper, a big traffic data processing framework using HBase to analyze the data of IMRS is proposed. The performance of the proposed framework is compared with the Oracle strategy on three different kinds of applications, and the experimental results show that the proposed strategy is with promising performance in computation latency.