Design Of Defective License Plate Recognition System

With the continuous development and progress of the economy,people's living standards are constantly improving,and the number of cars in various countries is also rising rapidly.The traffic problems caused by this have gradually attracted people's attention.In order to solve the increasingly serious traffic problems and improve the efficiency of transportation,the research and application of intelligent transportation systems are gradually increasing.As the unique symbol of a car,once the license plate is recognized,all the information of the vehicle can be obtained at a glance.Therefore,the research of license plate recognition is especially critical for intelligent transportation systems.As a crucial link in the intelligent traffic management system,the license plate recognition system has been widely used in many occasions such as highway electronic toll collection,access control and traffic monitoring.At this stage,the license plate recognition system recognizes that the accuracy rate of the license plate in the normal environment is relatively high,but the accuracy rate of the license plate recognition in the harsh environment or the pollution condition is relatively low.Aiming at this situation,this paper proposes a system for recognizing defaced license plates,which greatly improves the accuracy of recognizing defaced license plates,and plays a positive role in building intelligent road traffic management.The purpose of this design is to develop a license plate recognition system with high recognition rate,error rate,short time consumption,and can effectively determine the recognition of dirty license plates.The defaced license plate recognition system is mainly composed of three parts,including license plate image processing,license plate extraction and license plate recognition.The license plate image processing part mainly includes the determination of the license plate pollution situation,the repair of the license plate pollution,the Gaussian smooth filter processing,the grayscale processing,the adaptive segmented grayscale transformation and the histogram equalization.In the license plate extraction part,it is divided into two steps: generating a license plate area and screening a license plate area.Among them,the license plate area mainly includes edge detection of the license plate image,binarization processing,expansion and corrosion processing and image segmentation;the screening plate area is mainly composed of SVM classification and license plate image search and selection.In the final part of the license plate recognition,firstly,the tilted correction is performed on the filtered license plate image,and then the license plate characters are segmented using the OCR segmentation function,then the character features are extracted,and finally the artificial neural network recognition method and the character feature recognition method are used to recognize the license plate characters.In general,a relatively complete defaced license plate recognition system has been constructed.The process of designing the license plate recognition system involves a variety of image algorithms,which are implemented based on OpenCV,which improves the reliability and accuracy of the program;the dirty plate is judged based on the number of vertical projection jumps of the license plate character,and the damaged plate is repaired using the Criminisi algorithm;In the process of license plate character recognition,the artificial neural network method combined with character feature recognition method is used to classify the numbers,letters and Chinese characters in the license plate for training and recognition,making the system's license plate recognition rate ideal.Through testing and verification of the system,it is found that this system is ideal in terms of accuracy of license plate positioning and character recognition compared to similar systems.