Research On License Plate Recognition Technology In Video Image

License plate recognition technology is an important part of intelligent transportation,and plays an important role in such important occasions as urban road monitoring,electronic non-stop charging and community access control.Based on the theory of video image processing,studies the method of license plate recognition technology,and mainly improves the algorithms of license plate location,license plate character segmentation and license plate character recognition.In the license plate detection part,in view of the low detection rate of the traditional license plate.A method for cascade adaboost license plate detection based on HSV color model and multi-block local binary patterns(MB_LBP)is presented to realize fast and accurate license plate detection and recognition.Firstly,the license plate image is transformed from RGB color space to HSV color space,and the ratio of the blue pixels to the total pixels of the license plate is counted to construct the first class strong classifier.Then,the MB_LBP feature is extracted from the license plate character samples,and the feature selection and the classifier training are carried out by using the Adaboost classifier training method.Finally,a new license plate detection algorithm is formed by using the Cascade structure detection method.In the license plate character segmentation part.A license plate character segmentation method based on Blob analysis algorithm is proposed.Firstly,the blob analysis algorithm combined with prior knowledge of license plate characters for rough segmentation.Secondly,using vertical projection method to achieve the fine segmentation of the characters.Experimental results show that the algorithm can effectively solve the problem of disjoint Chinese characters segmentation and character stickiness,and has high practicability.In the license plate character recognition part,Aiming at the problem that the traditional BP neural network has a slow convergence rate and easy to fall into the local optimum.A license plate recognition algorithm based on improved PSO-optimized BP neural network is presented.Firstly,the sobel edge detection and harris corner feature extraction are taken as the input of BP neural network,and adaptive mutation operator is proposed in pso to change positions of the particles which plunged in the local optimization.The modified PSO was used to optimizethe weight and learning strategy are improved.Then the improved PSO algorithm is used to optimize the weights and thresholds of the BP neural network to minimize the fitness value.Compared with the traditional BP algorithm,the algorithm has higher convergence speed and recognition rate.Using the improved license plate detection,character segmentation and character recognition algorithms,the license plate recognition experiments in video images were carried out.The results show that the proposed algorithm has better applicability than the similar algorithms.