Design And Implemetation Of Bus Real-time Human Traffic Statistics System Based On Embedded System

The intelligent scheduling is the key part of the Intelligent Transportation System(ITS),It plays an important role in alleviating urban traffic jam,improving the utilization rate of public resources,reducing the operating costs of public transport system and facilitating public travel.And the real-time,accurate bus passenger-flow information is the most important data for intelligent scheduling.In this paper,I transplanted the Linux operating system,OpenCV(Open Source Computer Vision Library),Qt library to the ARM embedded platform,and realized the design and implementation of the bus human traffic statistical system.Finally,the statistical information is sent to the database with the IP address specified by GPRS.In this paper,introduced the installing of the Linux kernel,CRAMFS root file system to the ARM platform,the transplantation of the OpenCV,Qt library,then designed the traffic statistics system for the scene that the passengers getting on/off the bus.In this system,according to the difference of the crowd density,we use different methods to statistic the human traffic.In crowded occasion,passengers get too closed to each other,the foreground image at this time will be an area of mutual adhesion,and the segmentation or detection of moving targets cannot be completed.In this paper,Hough Circle Transform,centroid coordinates,colors are used as features to detect and track the passengers.And the method is tested at the end of thepaper.The results show that,the method provided in this paper solves the problem that the moving objects cannot be split in the high-dense traffic occasions,and it has good real-time and robustness,makes a supplement for the statistics of real-time bus traffic,the practical significance is very important.