The Research And Implementation Of Traffic Flow Based On Motion Detection Technology

With the development of city road, the Intelligent Transportation System(ITS), which has a brilliant future for the expansion of transportation, was used commonly in the application of Transportation Management System. At beginning, the traffic flow data was collected by non-motion monitor technique,Which was unstable and biased. Therefore, to develop a robust and stable traffic flow monitor system is very important and useful. It is not only provides the accurate traffic flow data for the transportation management department, but also makes the transportation management more intelligent, and more efficient.This paper is to illustrate the four modules of traffic flow management system, based on motion detection technique:Adaptive Background Modeling,Background Subtraction Process, Virtual Coil Technique and Traffic Flow Management. The first three stages were introduced in this paper, since the whole system was based on them. The key point is using motion detection and Virtual Coil technique to make preparation for the next modules. It was stated as follow:1.Since it was difficult to detect the target in the motion detection process when the backgrounds vary constantly, there is a new method to figure out this problem. The method uses adaptive background modeling to deal with the difficulty of target locating problem, concerning all kinds of interference effect, and combining different background modeling methods.After the collection of vehicle video and the modeling of adaptive background, the Background Subtraction Process was used to separate the valid target. Then, by digital image processing, gray scale setup, image enhancement, image binaryzation, edge contour detection, shadow elimination with Morphology, ROI target segmentation, the valid target cells were extracted from the video frame sequence. The result shows that the target was successfully located at a probability of94%, given the existing conditions.2.Based on the complexity of the process of a vehicle passing the virtual coil and the knowledge of digital image processing, a new detection technique,improving the degree of filling with edge contour detection, was introduced in the third modules. The idea is:Firstly, locate the length and width of the target vehicle contour with motion detection technique. Secondly,build a program to collect the information from the vehicle passing through the virtual coil. Finally, use the area of vehicle contour to fill the area of virtual coil, also calculate the proportion. The accuracy rate of90%was got, comparing with the conventional method.Besides that, these three modules will decrease the computation cost of whole traffic flow management system in the next procedure. Also, it increases the stability, intelligence and accuracy of intelligent transportation management.