Font Size: a A A

Research On The Key Technologies Of The License Plate Recognition System

Posted on:2013-10-27Degree:MasterType:Thesis
Country:ChinaCandidate:Y ChenFull Text:PDF
GTID:2248330374490680Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
The License plate recognition system is the core technology of the IntelligentTransportation Systems,which to solve the increasingly serious road traffic management, andcan be used for the detection of traffic flow, traffic crossing at the highway toll points, airports,ports, car parks, residential, can be easily effective vehicle monitoring and management,hasbroad application prospects, has been of concern. Its research mainly deals with a number ofimage processing, pattern recognition and artificial intelligence technology, the image dataacquisition and analysis, to obtain useful information, so as to achieve a high intelligentmanagement.In this paper, the license plate recognition system is divided into the following four parts:pretreatment of the license plate, license plate location, character segmentation, characterrecognition. Artificial neural networks with particle swarm optimization was used for thelicense plate character recognition,100license plate images was used as an experimental,identification rate of92%with better recognition effect. The following three aspects weremainly studied in this paper.Ⅰ. For the current algorithm in license plate location and tilt correction part of thedeficiency, the Radon transform method for license plate tilt correction, andaccording to the texture characteristics of plate area, positioning strategy based onthe connected domain and texture features, greatly improve the positioning effect ofthe license plate;Ⅱ. The advantages and disadvantages of the method of license plate recognition basedon neural networks, the PSO algorithm optimized artificial neural networks, particleswarm optimization algorithm has to rely on experience with fewer parameters, andconvergence speed advantages of training the neural network threshold weights, thesimulation shows that the improved BP neural network algorithm has fasterconvergence speed, significantly reducing the network training,the use of time;Ⅲ. On the basis of theoretical studies, this article by VC++6.0on the various aspects ofthe algorithm for simulation and verification, compare the pros and cons of eachmodule algorithm to find the optimal strategy using matlab for the development outof the bus, and then under the policy brand recognition systems, and simulationresults.
Keywords/Search Tags:License Plate Recognition, License Plate Location, Character Segmentation, Particle Swarm Optimization Algorithm, Neural Network
PDF Full Text Request
Related items