| Intelligent vehicle flow detection technology is one of the key technologies of intelligent transportation system; it can detect the traffic parameters such as vehicle type, vehicle speed, traffic flow, road occupancy, etc. According to these parameters, judge the traffic condition of the road, and optimizate the weak link in the traffic. At present, the traffic signal lamp switch is fixed by the manual one time, but the actual response of the traffic conditions of the parameters are dynamic, it can not change traffic lights status according to the actual traffic flow, and will reduce the utilization rate of traffic resources necessarily. In order to avoid the situation, reduce the vehicle’s stopping time and improve the vehicle passing rate, it necessary to control the traffic lights and green light switch time dynamicly according to the traffic flow. Moreover, intelligent transportation system is a highly integrated system, need to return the traffic situation data of each intersection to the traffic control and analysis comprehensively, provide accurate data and support for the vehicle’s navigation.In view of the problems existing in the practical application, it is needed to study a kind of intelligent traffic signal control system based on machine vision. This system is one kind of embedded machine vision application device, the traffic flow data of the road vehicle can be obtained, through the processing and analysis of the real-time video images, then control traffic signal device intelligently, while providing data source for advanced application. In addition, the other goal of this research is to realize the embedded platform of machine vision application with low coupling degree of a soft hardware system, minimize the coupling degree between the machine vision algorithm and hardware, the benefit of doing like this are: First, we can maximize the use of existing development resources, reduce the difficulty and complexity of development, accelerate the formation of products; Second,it can shorten the product development cycle and reduce development costs,development difficulty and complexity, due to rich the software and hardware resources; Third, the software and hardware of the system adopts the module development, which can be integrated with the actual application requirements, and it can enhance the scalability of the system; Fourth, the method of this research can be used to transfer the mobile phone and other mobile terminals, which can be used to open up more applications and markets.This study has innovative points as bellow:(1) Explore one kind of the software and hardware modules realize method different with traditional way of the machine vision algorithm application in the embedded system;(2) Using the latest ARM processor, realizing the application of machine vision graphics image processing well, so, the method has a good flexibility, portability and scalability, reducing cost of R&D and production, and it can be used in civil system, especially in mobile terminal area;(3) Realized transplantation of the famous ACE communication engine and graphics image library OpenCV to the embedded system successfully, which makes it possible to apply the machine vision application by the ideal of the Low coupling degree theory of software and hardware;(4) Constructed a set of embedded machine vision application platform successfully, it can transplant machine vision algorithm module very convenient and Formation of embedded machine vision production rapidly.(5) Successfully implemented in the embedded system based on OpenCV machine vision vehicle flow detection algorithm, which includes the use of statistical methods to achieve road background reconstruction, the use of background subtraction and maximum inter class variance method to achieve the vehicle target extraction, the use of virtual loop method is realized the vehicle count. |
PDF Full Text Request