The Research Of Vechile Detection And Identification In Intelligent Transportation System

With the development of the economy, intelligent traffic theory arouse the people's attention. The topic of this article mainly discussed the vehicle video processing, including image gathering, image pre-processing, vehicles detection and recognition, and so on.Firstly, this article introduced the design of the entire vehicles image gathering system in intelligent traffic system. Through man-machine interface way it made use of Visual C + + 6.0 integrated development tool for Mortimer Video Technology to open up Inc. SDK2000 image gathering card. During the course, the system mainly realized the preparation of the equipment installed, video settings, video preservation, video compression, video transmission and so on. Then, to do with the actual image include image de-noising, contrast enhancement, and so on. These preprocesses were ready for further image analysis processing.In the vehicles detection algorithm research, a novel fusion detection approach based on both the inter-frame subtraction and background subtraction was proposed. It overcame the single classical vehicles detection method, such as the high undetected rate. Experimental results indicated that, this fusion method was superior to both inter-frame subtraction and background subtraction method, the detection precision was high and mistaking rate greatly reduced. During the course, the shadow elimination algorithm had been studied, the disturbance was further removed and the examination precision increased.For the vehicles recognition, edge detection algorithm was commonly used to carry on the vehicles outline extraction, but the smooth continual contour line cannot be obtained .Therefore this paper introduced the mathematics morphology method, it could carry on the vehicles outline extraction effectively. Then, calculating the vehicles' parameter characteristic. Finally, using the Fisher sorter to carry on the distinction and realize the three kinds of vehicles recognition, including the large car, the small car, and the super car.