Study On Altered License Plate Recognition Method

License plate recognition technology plays an important part of the safe city and smart city.It is mainly used in video surveillance systems for highway vehicles,parking lot management systems,and vehicle escape monitoring systems.In the vehicle management system,the vehicle license plate is the number plate of the car's identity.Accurate identification of this is the key to the normal operation of the Intelligent Traffic System(ITS).The license plate character recognition is the decisive link of the license plate recognition system.Deliberately modifying the license plate will interfere with the normal operation of the intelligent traffic system,affect the road traffic safety,and disturb the public order.Therefore,it is of great significance to carry out research on changing the license plate recognition.The altered license plate seriously affects the inherent characteristics of the license plate in terms of texture,color,character arrangement structure and geometric dimensions,which makes the traditional license plate recognition technology and conventional monitoring system unable to work normally.There are two methods to modify the license plate: simple alteration and complicated alteration.This paper has carried out relevant research on different methods of correction.This paper proposes an improved recognition algorithm for template character library for simple modification.In view of the complicated modification methods,this paper starts with the retro-reflection characteristics of license plate materials,designs an active-passive polarization imaging system to obtain the altered license plate polarization image,and combines image processing algorithms.To accurately identify the altered license plate,the specific research work is as follows:1.For the situation where the license plate is simply altered,this paper designs an improved template matching method.By refining the traditional character library and increasing the subcharacter library,a more detailed character library is built than the traditional character library.When the discrimination result of the altered portion is discriminated,the altered character is extracted and re-recognized.The experimental results show that in the case of the failure of the traditional license plate recognition algorithm,this algorithm can accurately identify the altered license plate.2.For the situation where the license plate is complicated to be altered,because the color of the altered material is very closed to the license plate and finely altered,conventional imaging equipment cannot perform normal discrimination andmisidentification occurs.Based on the retro-reflective characteristics of the license plate material,this paper designs a altered license plate recognition method based on passive light polarization.Through the use of passive light polarization imaging system to obtain the license plate polarization image,combined with the image processing algorithm to segment the altered part and the original license plate number.Experimental results show that this method can identify characters that are realistic,visually indistinguishable from the imaging device,and accurately extract the original license plate number and the altered portion.3.For the limitations of passive light polarization imaging system to control illumination,this paper designs a method based on active light polarization imaging license plate,which can identify the altered license plate at night or in low light environment.The experimental results show that the method can accurately and effectively identify the original license plate number of a license plate in a low-light environment and efficiently separate and display the altered portion.