基于深度学习的端到端车牌信息检测与识别

End-to-end license plate information detection and recognition based on deep learning

在智能交通系统中,车牌检测与识别是实现车辆管理和监控的重要环节,传统的车牌识别算法在定位准确性、识别效率和实时性方面存在诸多问题。文章提出一种基于深度学习的端到端车牌检测与识别系统,通过YOLOv8n算法进行车牌定位,利用空间变换网络(STN)对倾斜车牌进行矫正,并通过PaddleOCR进行字符识别。实验结果表明,该方法在CCPD数据集上的识别准确率达到98.5%,单张图像的检测识别时间控制在80ms以内,优于传统方法。

In intelligent transportation systems,license plate detection and recognition are important links in achieving vehicle management and monitoring.Traditional license plate recognition algorithms have many problems in terms of positioning accuracy,recognition efficiency,and real time performance.This article proposes an end to end license plate detection and recognition system based on deep learning.The YOLOv8n algorithm is used for license plate localization,Space Transform Network(STN)is used to correct tilted license plates,and PaddleOCR is used for character recognition.The experimental results show that the recognition accuracy of the proposed method on the CCPD dataset reaches 98.5%,and the detection and recognition time of a single image is controlled within 80ms,which is better than traditional methods.