摘要
随着汽车保有量日益增多,车牌识别已经成为智能化交通系统不可分割的一部分,在道路交通控制方面发挥着重要作用。由于实际交通环境的复杂性,传统的车牌识别算法易受到光线明暗不均、天气恶劣多变、倾斜角度大等诸多因素的干扰,具有很大的局限性。针对这一问题,本文提出了一种基于深度学习的复杂场景下车牌识别算法。从车牌识别的一般流程出发,设计基于HOG+SVM目标检测定位算法,实现对车牌的定位与筛选,最后利用CNN卷积神经网络对车牌字符实现识别。实验结果表明,在复杂多变的应用场景下,本文所提出的基于深度学习的车牌识别算法精度高、鲁棒性强。
With the increasing number of cars,license plate recognition has become an integral part of the intelligent transportation system and played an important role in road traffic control.Due to the complexity of the actual traffic environment,the traditional license plate recognition algorithm is susceptible to the interference of many factors,such as uneven light and dark,bad and changeable weather,large inclination angle and so on,which has great limitations.To address this problem,the article proposes a license plate recognition algorithm in complex scenes based on deep learning.Starting from the general process of license plate recognition,a target detection and location algorithm based on HOG+SVM is designed to realize the location and screening of license plates.Therefore,CNN convolutional neural network is used to realize the character recognition of license plates.The experimental results show that in complex and changeable application scenarios,the license plate recognition algorithm based on deep learning proposed in this paper has high accuracy and strong robustness.
作者
杨金鑫
颜湘炎
王子宁
YANG Jinxin;YAN Xiangyan;WANG Zining(School of Automotive and Traffic Engineering,Jiangsu University,Zhenjiang Jiangsu 212013,China)
出处
《智能计算机与应用》
2023年第3期33-38,共6页
Intelligent Computer and Applications
关键词
车牌识别
深度学习
复杂环境
SVM
卷积神经网络
license plate recognition
deep learning
complex scene
Support Vector Machine
convolutional neural network
