Vehicle Identification Technology Research And Application

The growing development of intelligent transportation systems, makes use of the vehicle information for vehicle recognition techno-logy growing and concern, the vehicle identification and classification technology is based on digital image study, it is an Applicability technology of vehicle brand and model information is identified by the image processing and automatic identification method. Vehicle logo recognition technology as an important aspect of intelligent transportations systems, its application is also received more and more attention of researchers. The car standard Automatic Identification System Consist primarily of two key technology vehicle location and identification. The vehicle-logo positioning is a typical target detection and location technology. Its main purpose is to deck the capture image whether or not we have needed vehicle-logo information. Because of currently the traffic bayonet mostly is one-way camera, Automobile rear information automatic identification system did not receive due diligence, however right now some cities and traffic road have provide two-side camera, so the automobile rear information automatic identification system will be our next target on ITS. Automobile rear information automatic identifications system also contain two key technology about rear information location and identification. With the car logo recognition and rear information recognition technology research deepens, intelligent transportation system will eventually reach the unattended goal.This article use traffic mount camera shooting images as study object, use image processing and pattern recognition technology combined with the mathematical knowledge of geometry to do locate and identify the car standard and rear vehicle classification algorithm Exploration and Research.in this paper we innovation propose use Hough Transform and corner detection method on location car-logo; innovation propose use identification car rear information to identify car type. And did lots of experiment to prove our methods is practicable. In this paper we mainly do follow works:1) Established a standard library of car models. In image processing and pattern recognition, sample template is essentially but now there is not exit a complete vehicle type standard library, in order to do a solid foundation for our study, we did a lot work to established a standard library. In this library we divided all vehicle in to three type: family-use, truck and bus. And subdivision this three type into more classification like family-use to one box car, hatchback, the sedan, SUV, MPV and other types. Every vehicle has three standard library, contains vehicle’s linear geometric features, rear information and front positive messages. The vehicle type also have a car-logo library.2) Research and application of car-logo location and identify. In the vehicle-logo location aspect, we propose a method about multiple rough localization and final precise localization. Use a series of a priori knowledge about vehicle-logo’s position to do rough localization so we can get an approximate range about vehicle-logo’s position. In precise localization we propose some new methods to do vehicle-logo precise localization—Hough Transform and corner detection. We use Hough Transform do precise localization on like round vehicle-logo、oval vehicle-logo and so they have a closed contour. Use corner detection method do precise localization on those vehicle who do not have a closed contour. On vehicle-logo identify aspect we use a pattern recognize method based on Hu invariant moments,3) The localization and identify of vehicle rear information. According to the information about the characteristics of the rear of vehicle, we proposed a subtraction method to locate vehicle rear information so that we can eliminate interference information. First use of license plate location method to locate the license plate area, then pull in addition to the area from the rear area, the rest of the region information contained character is the vehicle of the car Details. In this paper, the character recognition method based on F-SVM to the rear of the vehicle information to identify the complete vehicle identification information through the rear of the experimental system. Verify the correctness of the use of the method of the rear information identification models.