Research On License Plate Recognition Algorithm Under The Fog-Haze Environment

The automatic vehicle license plate recognition technology is an important part of the intelligent transportation system.At present,that the image of vehicle license plate is processed is often used for us to obtain the information in vehicle license plate.In recent years,due to the destruction of the environment,the haze weather has gradually increased all over our country.Under the smoggy weather conditions,due to tiny particles in the atmosphere,the image of the vehicle license plate photographed by the camera will be degraded,resulting in a lower accuracy of vehicle license plate recognition..In order to improve the accuracy of vehicle license plate recognition in the smog environment,this paper first analyzes the steps of vehicle license plate recognition.Such as of image defogging,license plate location,license plate segmentation and license plate recognition.Based on this,we optimize the process of vehicle license plate recognition,and propose an optimized vehicle license plate recognition algorithm,by using the algorithm,we can easily improve the recognition effect and accuracy in the foggy weather environment.The main contents of this paper are as follows:1.Under the guidance of the dark primary color prior defogging algorithm,we proposes an improved method,which optimizes the transmittance by using adaptive median filtering,by using this method,we can improve the speed of image defogging.2.We proposes an improved vehicle license plate location method,which combined edge detection and machine learning method.Firstly,we grayed the atomized image,then we determined the upper and lower borders of the vehicle license plate by using the edge detection method,and the left and right borders of the vehicle license plate are divided by using prior knowledge algorithm;3.We binarized the image,and then the characters are segmented by using the vertical projection method,finally,we normalized the image,and base on this,the image was converted into a coarse feature matrix;4.The radial basis function neural network is used for identification.At the same time,in order to simplify the structure of the neural network and meet the accuracy requirements,we optimized the parameters of the neural network,by using this method,the accuracy is improved greatly.The vehicle license plate recognition algorithm proposed in this paper can not only improve the accuracy of vehicle license plate recognition in the haze environment,but also has little effect on the recognition speed in the sunny environment,which can be neglected.So it can greatly reduce the impact of weather factors in vehicle license plate recognition.