A Design Of The Entrance Guard System Based On The License Plate Recognition

With the quick development of modern science and technology,people's living standard is increasing,the vehicle has become one of the most people's quality of life,in daily life,people are more willing to choose to drive to work,which make pressure on the management of enterprises and factories.In order to solve the traffic management problems,this paper puts forward and designs an access control system based on license plate recognition.License Plate Recognition(LPR)technology is a comprehensive application of computer vision and pattern recognition technology.In this paper,license plate location,character segmentation and character recognition are studied.(1)The purpose of license plate location is to find the image location of the vehicle license plate,taking into account the multiple character changes of license content,and the different texture changes of the license suspension,pixel jump edges in the graph are detected through the edge detection;then according to the rectangular shape of the license plate,all edge points are morphological expansion processing,which will be within a certain range of the edge points of convergence,in addition to the area which does not conform to the characteristics of the connected domain of license plate,license plate image can be obtained;this paper also studies the positioning method based on pixel color,according to the distribution of blue pixel,in the candidate regions for statistics,get rid of the area which does not conform to the characteristics of the license plate,until you find the correct position.The method based on edge detection is better,and the calculation is less than the color location method,and the accuracy is high.(2)Character recognition.In this paper,pixel based projection,character block and cluster analysis based cutting method are introduced.Pixel projection method is the most simple,fast program and less amount of calculation.The new energy vehicle license plate is currently being studied for cutting algorithm,due to the new energy plate background for green gradient,this paper are studied from the perspective of green component RGB license plate license plate cutting,the experimental results obtains good effect diagram,and shows that the method can be performed.(3)Character recognition,template matching and neural network are used in character recognition,the results shows that the two methods can get higher correct recognition,but the recognition method of neural network owns strong anti-jamming ability,in compare with the experimental results of BP neural network and RBF neural network analysis,the BP network is selected as a recognition method of access control system.Finally,data storage module and door brake control module are added to license plate recognition technology,a complete access control system are designed,the function design and hardware adaptation are included in each module.In the test results,the system designed in this paper can locate,cut and identify the license platevery well,which indicates that the scheme and method with practical valueadopted are correct and feasible.