The Research Of Illumination?Distance And Angle In License Plate Recognition System

With the rapid development of China's automobile industry,the number of automobiles in China has increased from 240 million in 2018 to 250 million in 2019.In one hand,the increasing number of cars shortens people's travel time and it's convenient for people to travel;in another hand,it also brings great pressure to management departments.In order to reduce the workload of managing cars,the regulation of road traffic and the management of communities can be realized,it's necessary to manage traffic through the license plate recognition system.According to the image processing knowledge,this article designed a license plate recognition system for closed communities,schools and other places where vehicles need to pass at a low speed;for the needs of the subject the camera and light meter for the license plate recognition system were selected;according to the designed platform designed,a group of pictures were taken by controlling the relevant elements,and three aspects such as illumination ?distance and angle were explored.Firstly,this paper introduces two kinds of methods in license plate location.One is through grayscale transformation convert the color image to grayscale image,thought histogram equalization to enhanced the overall contrast of the image,median filter to removes noise,and then through edge detection,morphological processing and calculation of the length-width ratio of the license plate to complete the positioning of the license plate;the another is to complete the positioning of the license plate through HSI color space ? morphological processing and calculating the length-width ratio of the license plate.In the selection process,the second method is found to be more effective,with the increase of light intensity,the value of S in the HSI color space increase in a stepped manner,when the distance changes,without the interference of external light,the value of S gradually decreases with the distance between the license plate and the camera;under theinterference of external light,the value of S gradually increases with the distance between the light and the camera.After the license plate positioning is completed,Randon transform and Hough transform are used to correct the tilted license plate,and remove impurities on the license plate by morphological,when the length-width ratio of the license plate is x?(2.5,4),it is better to take the value range of angle [0°,30°].Though OCR technology of NI Vision Development Module in LabVIEW,the fixed license plate is segmented and trained so as to generate a template character library of.abc.The generated.abc file can be used to identify license plate characters on the license plate.Through the collection of experimental data and the verification of experimental data,it can be concluded that illumination,distance and angle will affect the license plate recognition system.It is better to apply the designed system to the reality word.