Research On The Key Technologies Of License Plate Recognition System Based On Image Processing

With the development of our national economy,there are more and more constructions in the expressway,urhan road,and parking area,ere.The requisition on the traffic control,safety management improves day by day.The intelligent transportation system(ITS)has already become the main direction of present traffic administration development,and vehicle License Plate Recognition(LPR) system as core of ITS,plays a very important role.It not only can widely be applied traffic flux detect,but also be applied traffic control and lead.It can also be applied to some fields where the vehicle breaks the laws,such as airport,seaport,communities'vehicle manage,no halted vehicle's fee,and vehicle's safe.So it has widely applied in the future.Basing on the current thing,the thesis will expound the course that searches and deals with image of the License Plate.LPR consists of three modules in general,those are:License Plate preprocessing, license plate location,character segmentation and character recognition.Algorithms of all modules related to LPRS is deeply studied and analyzed in this paper.And in this paper finally settle down a series of algorithms used in LPR based on our research. The following is the main innovating content of the thesis:During the pretreatment phase.At first,the thesis deals with the gray-scale and equalization of License Plate,then brings forward a new project method of noise reduction based on PDE.Although this method is a little complicated steps,it is a good solution to solve the conflicts of preprocessing smoothing and security details between borders.At the License Plate image location and pick-up.The thesis designs two methods of License Plate image location,locating License Plate which based on color feature integrate H information of HSI color space as characteristics of plates which based on traditional analysis methods.And locating License Plate which based on gray-hopping utilizes grey degree of characters on the plate to confirm location of plate.On the course of incising the License Plate character.Though there are many factors that cause the characters segmentation task difficult,such as plate frame, rivet,spacemark plate rotation,a new algorithm for character segmentation was presented.The characters were segmented with the method of connect components which based on the digital morphology.At the same time a valuable project for the divided procession of the characters of vehicle license plate was proposed.On the course of recognizing the License Plate character.Firstly,the thesis brings forward to pick up character of the License Plate character.Then we used the result to train the Backpropagation Network.Finally,we used the trained result to recognize the character and decided the result.According to the step that the thesis has taken,and the ways that we have taken broughtbefore we write this.The experiment shows the result is effective.