Research And Applicationof Vehicle Identification In Intelligent Transportation Systems

With the rapid economical development, Intelligent Transportation System is much moreoften applied. As an important part of Intelligent Transport System, the technology of vehicleidentification becomes a noticeable focus point. The technology is available in HighwayManagement Systems, Parking Management Systems, "electronic police" and other fields.The technology of vehicle identification does a good job in reducing traffic accidents andproviding a safe environment for vehicles, thus it has much practical value in the modernsociety. At present, many related products using vehicle identification came out, while theresearch never stops, there are still many scholars in doing further study and improvement.In this paper, the status of the Vehicle Identification Systems is introduced and theexisting identification techniques are analyzed. While a license plate recognition system basedon MATLAB is introduced. The technologies related to digital image processing is used astheoretical basis in this thesis; the object of this research is the license plate and the keytechnologies, such as vehicle plate locating, character segmentation and character recognition,are mainly discussed and researched.The system consists of four algorithm modules: image preprocessing, vehicle platelocating, character segmentation; character recognition. In image preprocessing,it uses astandardized algorithm such as gray and smoothing filtering; In vehicle plate locating module,rough localization algorithm is presented based on edge detection and hopping to frequencygray. First, the every region of possible license plate must be found out. Then in order toobtain the only license plate region, the above possible regions are distinguished through themethods of binary differential operator and horizontal and vertical direction of the grayprojection. Character segmentation module includes binary, skew correction, borderidentification and character cut. The characters are segmented by combining the vertical projection of binary images with a priori knowledge .While the character recognitionalgorithm completes character recognition by using the template matching method accordingto the characters characteristics.According to the algorithms mentioned above, a simulation system is established usingM language of MATLAB for the whole system. The vehicle license plate image photo isidentified in the system. The experimental results show that the system can achieve the licenseplate recognition effectively and provide a good theoretical basis for manufacturing practicalproducts in the future.