Research on License Plate Location Algorithm Based on YOLOv5

Abstract

License plate location is a key part of license plate recognition, and its positioning accuracy seriously determines the final result of license plate recognition. Firstly, the traditional license plate location methods are compared and analyzed. Secondly, in order to solve the localization problem in low resolution and multi-vehicle environment, a license plate method based on YOLOv5s was proposed by using deep learning image recognition technology, and data enhancement was introduced to improve YOLOv5s. Finally, the YOLOv5 algorithm is applied to real vehicle images under complex environmental conditions for experimental verification, and the results show that the accuracy is 99.12%. The experimental results show that the license plate location model based on YOLOv5S can effectively solve the influence of illumination, image quality and other factors on license plate location, improve the efficiency and accuracy compared with the traditional license plate location method, the algorithm has good robustness and fast calculation speed.