License Plate Recognition Key Technology Related Algorithms Research And Implementation

Image processing and recognition of vehicle license plate(LPR)system is an important computer image processing system,it can effectively ingest contains images and identify the license plate number plate,in many occasions,such as highway automatic charging,intelligent transportation,the suspect vehicle location and so on are important.In this paper,the license plate images of foggy weather and light intensity over bright and dark are collected,and the collected images are processed step by step.Finally,the license plate information is identified.License plate image preprocessing: first,the color images are removed by fog,then gray and histogram enhancement,low-pass filtering and edge image processing are carried out.License plate location: through the previous step to extract the edge map,through row by line scan,and combined with the license plate length and width ratio and other geometric features,locate the license plate position.Character segmentation: the extracted region of the license plate is iterated,two values are processed,and the inclination is adjusted at the same time.Each character is determined by judging the connected domain.Character recognition: segmentation of the characters obtained,resizing and refining,extract15 features,and compared to the template library,identify each character.The experimental data show that after adopting a series of methods proposed in this paper,the accuracy of license plate recognition can reach more than 85%.