Traffic Road Traffic Detection Technology Based Virtual Area

With the development of computer hardware and software technology, video traffic detection technology has been tremendous development, at the same time, it can be easily installed and maintained, which can provided a large amount of information, is the magnetic coil and other traditional traffic detection methods can not matched, so it has been widely used and developed, has become a key factor in intelligent transportation system.OpenCV is a function library which used by open source graphics process,it is powerful, can run on multiple platforms, and can meet the various needs of video traffic identification, allows researchers to extricate from the preparation of the underlying algorithm, t focused on the application development, which saves time and effort, and to ensure the stability and security of the application programs. Various traffic detection methods existing have some problem in installation and maintenance, test conditions and some required other conditions, and video detection is an alternative, not only can all-time detection, and the quantity of information extraction, high flexibility.This paper firstly introduces the development status in China and abroad as well as existing problems, then introduces some important OpenCV functions and data structures, analyses the key technology of digital image processing. After analysis and comparison of the existing vehicle detection method to study the difference of traffic flow detection method based on background edge subtraction, the background information of the vehicle as detecting characteristic, obtained data to extract and update the background information has been, getting as much data as less affected by the environment, on the basis to calculate vehicle flow detection method based on virtual coil.Secondly, analyze the problems in the process of detection, removing the image noise by minimum pixel denoising and minimum number denoising, try to reduce the impact of external noise on the vehicle identification. The vehicle shadows are treated separately in RGB space and HSV space, use the characteristic of shadows in the image, achieving the purpose of eliminating shadow. In addition, the vehicle adhesion phenomenon caused by the distance is too near, and processed using the maximum time interval threshold method, to solve the vehicle due to problems caused by fault gray minimum frame interval threshold method, based on the use of the OpenCV function with the Qt design video vehicle detection system. Finally, make a summary of the work, make some investigating according to the improved method of deficiencies and shortcomings.