基于卷积神经网络的车牌识别研究  

License Plate Recognition Based on Convolutional Neural Networks

作　　者：齐佳鑫 张志华[1] 付金尉 贺紫菡 司志广 王艺雄 QI Jiaxin;ZHANG Zhihua;FU Jinyu;HE Zihan;SI Zhiguang;WANG Yixiong(School of Electronics and Information Engineering,Liaoning University of Science and Technology,Anshan 114051,China)

机构地区：[1]辽宁科技大学电子与信息工程学院,辽宁鞍山114051

出　　处：《科技创新与生产力》2024年第4期121-124,共4页Sci-tech Innovation and Productivity

基　　金：辽宁科技大学大学生创新创业训练计划项目(X202310146303)。

摘　　要：为了满足交通管理和安全监控需求,针对车牌识别技术在智能交通系统中的重要性,本文旨在运用卷积神经网络实现汽车牌照的识别,首先通过对采集到的图像进行滤波降噪等预处理,运用Sobel算子进行边缘检测;其次对图像进行形态学运算,利用轮廓检测确定出车牌的位置;然后采取边缘检测的方法进行字符分割;最后运用卷积神经网络对车牌字符实现识别。模型测试结果表明,车牌识别系统汉字识别的准确率为91.4%,数字和字母共同识别的准确率为95.5%。In order to meet the demand for traffic management and safety monitoring,aiming at the important role of license plate recognition technology in the intelligent transportation system.This paper aims to use convolutional neural networks to recognize car license plates.Firstly,pre-processing such as filtering and denoising is performed on the collected images,and Sobel operators are used for edge detection;Secondly,perform morphological operations on the image and use contour detection to determine the position of the license plate;then,edge detection method is adopted for character segmentation;finally,convolutional neural networks are used to recognize license plate characters.The model test results show that the accuracy of Chinese character recognition in the license plate recognition system is 91.4%,and the accuracy of jointly recognizing numbers and letters is 95.5%.

关 键 词：神经网络 模型 车牌识别 车牌识别系统 边缘检测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]